Use of Geosynthetics for Separation and Stabilization in Low-Volume Roadways http://www.virginiadot.org/vtrc/main/online_reports/pdf/20-r8.pdf EDWARD J. HOPPE, Ph.D., P.E. Associate Principal Research Scientist Virginia Transportation Research Council M. SHABBIR HOSSAIN, Ph.D., P.E. Associate Principal Research Scientist Virginia Transportation Research Council AUDREY K. MORUZA Senior Research Scientist Virginia Transportation Research Council CHAZ B. WEAVER, P.E., CPEM District Materials Engineer Virginia Department of Transportation

Final Report VTRC 20-R8

Standard Title Page - Report on Federally Funded Project 1. Report No.: 2. Government Accession No.: 3. Recipient’s Catalog No.: FHWA/VTRC 20-R8

4. Title and Subtitle: 5. Report Date: Use of Geosynthetics for Separation and Stabilization in Low-Volume Roadways

October 2019 6. Performing Organization Code:

7. Author(s): Edward J. Hoppe, Ph.D., P.E., M. Shabbir Hossain, Ph.D., P.E., Audrey K. Moruza, and Chaz B. Weaver, P.E., CPEM

8. Performing Organization Report No.: VTRC 20-R8

9. Performing Organization and Address: Virginia Center for Transportation Innovation and Research 530 Edgemont Road Charlottesville, VA 22903

10. Work Unit No. (TRAIS): 11. Contract or Grant No.: 109960

12. Sponsoring Agencies’ Name and Address: 13. Type of Report and Period Covered: Virginia Department of Transportation 1401 E. Broad Street Richmond, VA 23219

Federal Highway Administration 400 North 8th Street, Room 750 Richmond, VA 23219-4825

Final 14. Sponsoring Agency Code:

15. Supplementary Notes: This is an SPR-B report. 16. Abstract:

The implementation of geosynthetics in road construction in the United States dates back to the 1970s. Although the benefits of these materials are widely recognized, actual design and construction practices are not universal. Guidelines for use are often developed locally based on specific field experience.

Currently, the Virginia Department of Transportation (VDOT) provides specifications (but no design guidance) for only one function of subgrade geotextile: stabilization. Pavement layer separation is not explicitly addressed by VDOT. This study reviewed corresponding practices at other transportation agencies, focusing on subgrade and pavement applications. In addition, observations gathered from field testing of two known sites with geosynthetics in Virginia were analyzed. Finally, life cycle cost analyses (LCCAs) were performed on two common secondary road designs, each with the options of separator geotextile and preventive maintenance, to estimate comparative costs under three different rates of subbase contamination by fines migrating from the subgrade.

The LCCA employed actual VDOT cost data and conceptual pavement layer deterioration curves that are based on AASHTO curves augmented by differentiated subbase deterioration resulting from three hypothetical contamination rates. Unlike the current LCCA used by VDOT’s Materials Division, the LCCA model developed in this study explicitly recognizes the fact of potential subbase layer deterioration because of contamination by subgrade fines, comporting with the widely recognized function of geotextile separator. The results indicate that separator geotextile imparts long-run cost-effectiveness relative to pavements without geotextile separators at contamination rates of 0.1 in per year and greater. At contamination rates of 0.05 in per year and lower, the addition of separator geotextile is not cost-effective, although the results vary slightly depending on the treatment of estimated pavement design life remaining at the end of the analysis period.

The study recommends that VDOT revise its current specifications to include subgrade separation and subgrade

stabilization geosynthetics as separate and distinct pay items, and it provides suggested specification language to effect this change. Guidelines to implement these changes are also enumerated. 17 Key Words: 18. Distribution Statement: Geosynthetics, geotextile, geogrid, roadway, separation, stabilization, LCCA

No restrictions. This document is available to the public through NTIS, Springfield, VA 22161.

19. Security Classif. (of this report): 20. Security Classif. (of this page): 21. No. of Pages: 22. Price: Unclassified Unclassified 93

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

FINAL REPORT

USE OF GEOSYNTHETICS FOR SEPARATION AND STABILIZATION

IN LOW-VOLUME ROADWAYS

Edward J. Hoppe, Ph.D., P.E.

Associate Principal Research Scientist

Virginia Transportation Research Council

M. Shabbir Hossain, Ph.D., P.E.

Associate Principal Research Scientist

Virginia Transportation Research Council

Audrey K. Moruza

Senior Research Scientist

Virginia Transportation Research Council

Chaz B. Weaver, P.E., CPEM

District Materials Engineer

Virginia Department of Transportation

In Cooperation with the U.S. Department of Transportation

Federal Highway Administration

Virginia Transportation Research Council

(A partnership of the Virginia Department of Transportation

and the University of Virginia since 1948)

Charlottesville, Virginia

November 2019

VTRC 20-R8

ii

DISCLAIMER

The contents of this report reflect the views of the authors, who are responsible for the

facts and the accuracy of the data presented herein. The contents do not necessarily reflect the

official views or policies of the Virginia Department of Transportation, the Commonwealth

Transportation Board, or the Federal Highway Administration. This report does not constitute a

standard, specification, or regulation. Any inclusion of manufacturer names, trade names, or

trademarks is for identification purposes only and is not to be considered an endorsement.

Copyright 2019 by the Commonwealth of Virginia.

All rights reserved.

iii

TABLE OF CONTENTS

ABSTRACT .................................................................................................................................... v

INTRODUCTION .......................................................................................................................... 1

PURPOSE AND SCOPE ................................................................................................................ 2

METHODS ..................................................................................................................................... 3

Overview ..................................................................................................................................... 3

Literature Review........................................................................................................................ 3

Determination of State of the Practice in Virginia ...................................................................... 4

Analysis of Test Sites With Geosynthetics in Virginia .............................................................. 4

Development of New LCCA Method ......................................................................................... 8

RESULTS ..................................................................................................................................... 15

Literature Review...................................................................................................................... 15

Performance of Sites With Geosynthetics in Virginia .............................................................. 32

Life Cycle Cost Analysis .......................................................................................................... 41

DISCUSSION ............................................................................................................................... 50

CONCLUSIONS........................................................................................................................... 56

RECOMMENDATIONS .............................................................................................................. 58

SUGGESTIONS FOR FURTHER RESEARCH ......................................................................... 59

IMPLEMENTATION AND BENEFITS ..................................................................................... 59

Implementation ......................................................................................................................... 59

Benefits ..................................................................................................................................... 59

ACKNOWLEDGMENTS ............................................................................................................ 60

REFERENCES ............................................................................................................................. 60

APPENDIX A: SURVEY OF VDOT RESIDENT ENGINEERS ............................................... 69

APPENDIX B: LAYER DETERIORATION CURVES AND MAINTENANCE SCHEDULE

DETERMINATION PROCESS SUPPORTING THE LCCA ..................................................... 73

APPENDIX C: EXAMPLE OF MAINTENANCE SCHEDULE DETERMINATION

PROCESS: NO GTX PAVEMENT OPTION WITH AND WITHOUT PM .............................. 81

APPENDIX D: EXAMPLES OF PAVEMENT MAINTENANCE SCHEDULES WITH AND

WITHOUT PREVENTIVE MAINTENANCE ............................................................................ 83

APPENDIX E: PROPOSED SPECIAL PROVISION FOR SUBGRADE SEPARATION AND

SUBGRADE STABILIZATION GEOTEXTILES ...................................................................... 85

iv

v

ABSTRACT

The implementation of geosynthetics in road construction in the United States dates back

to the 1970s. Although the benefits of these materials are widely recognized, actual design and

construction practices are not universal. Guidelines for use are often developed locally based on

specific field experience.

Currently, the Virginia Department of Transportation (VDOT) provides specifications

(but no design guidance) for only one function of subgrade geotextile: stabilization. Pavement

layer separation is not explicitly addressed by VDOT. This study reviewed corresponding

practices at other transportation agencies, focusing on subgrade and pavement applications. In

addition, observations gathered from field testing of two known sites with geosynthetics in

Virginia were analyzed. Finally, life cycle cost analyses (LCCAs) were performed on two

common secondary road designs, each with the options of separator geotextile and preventive

maintenance, to estimate comparative costs under three different rates of subbase contamination

by fines migrating from the subgrade.

The LCCA employed actual VDOT cost data and conceptual pavement layer

deterioration curves that are based on AASHTO curves augmented by differentiated subbase

deterioration resulting from three hypothetical contamination rates. Unlike the current LCCA

used by VDOT’s Materials Division, the LCCA model developed in this study explicitly

recognizes the fact of potential subbase layer deterioration because of contamination by subgrade

fines, comporting with the widely recognized function of geotextile separator. The results

indicate that separator geotextile imparts long-run cost-effectiveness relative to pavements

without geotextile separators at contamination rates of 0.1 in per year and greater. At

contamination rates of 0.05 in per year and lower, the addition of separator geotextile is not cost-

effective, although the results vary slightly depending on the treatment of estimated pavement

design life remaining at the end of the analysis period.

The study recommends that VDOT revise its current specifications to include subgrade

separation and subgrade stabilization geosynthetics as separate and distinct pay items, and it

provides suggested specification language to effect this change. Guidelines to implement these

changes are also enumerated.

vi

FINAL REPORT

USE OF GEOSYNTHETICS FOR SEPARATION AND STABILIZATION

IN LOW-VOLUME ROADWAYS

Edward J. Hoppe, Ph.D., P.E.

Associate Principal Research Scientist

Virginia Transportation Research Council

M. Shabbir Hossain, Ph.D., P.E.

Associate Principal Research Scientist

Virginia Transportation Research Council

Audrey K. Moruza

Senior Research Scientist

Virginia Transportation Research Council

Chaz B. Weaver, P.E., CPEM

District Materials Engineer

Virginia Department of Transportation

INTRODUCTION

Geosynthetics are man-made materials designed for use in geotechnical applications with

the intent of improving engineering performance and facilitating construction. ASTM D 4439

(ASTM International [ASTM], 2018) defines geosynthetic as a “planar product manufactured

from a polymeric material used with soil, rock, earth, or other geotechnical engineering related

material as an integral part of a man-made project, structure, or system.” These materials can

serve various functions, including separation, stabilization, filtration, reinforcement, drainage,

and erosion protection. In road construction the primary applications involve subgrade

stabilization or reinforcement, separation between the subgrade and the unbound base material to

prevent intermixing of fines with aggregate bases, and filtration to allow the flow of water at the

subgrade-base interface. Recently there has been increased interest in the potential use of

geosynthetic interlayers in pavements to mitigate surface cracking.

Geosynthetic materials have been used with increased frequency in transportation

infrastructure since the1970s, but their incorporation into specifications and guidelines was

gradual. The first official mention of geotextiles in the Virginia Department of Transportation

(VDOT) pavement design method dates to 1995, when the Flexible Pavement Design Guide for

Primary and Interstate Roads in Virginia (VDOT, 1995) included the following guidance:

Geotextiles should be considered in lieu of the above [cement- and lime-stabilized subgrade and

aggregates] when the areas in question represent a relatively small amount of the subgrade soils.

This may prove to be more economical than the above alternatives in isolated cases.

2

This language was added to VDOT’s Pavement Design Guide for Subdivision and

Secondary Roads in Virginia (hereinafter “Pavement Design Guide”) (VDOT, 2000) and it

continued in nearly the same format in the latest 2018 edition (VDOT, 2018a), but with the word

“geotextile” replaced by “geosynthetic.” The current Pavement Design Guide (VDOT, 2018a)

does not identify the potential use of geotextile separators between subgrade soils and aggregate

layers and continues to reference Highway Subdrainage Design (Moulton, 1980), written at a

time when underdrains typically used a “filter layer” rather than a geosynthetic drainage fabric.

The VDOT Materials Division Manual of Instructions (MOI) includes some pavement

design guidance (VDOT, 2018b). However, the MOI does not include any specific

recommendations regarding subgrade stabilization options or separation of subgrade soils.

VDOT’s use of geosynthetics evolved over the years through field experience gained from

experimentation. Despite their proven benefits, to date no specific guidelines have been

articulated at VDOT to facilitate the systemic use of geosynthetics on road construction projects.

The creation of guidelines to incorporate a pay item into routine construction practice

implies that new incurred costs will be reasonable, but such an assumption should be tested, to

the extent possible, under current VDOT costs. This study employed actual VDOT cost data and

conceptual pavement layer deterioration curves that are based on AASHTO (American

Association of State Highway and Transportation Officials) curves augmented by differentiated

subbase deterioration resulting from three hypothetical contamination rates. Unlike the current

life cycle cost analysis (LCCA) used by VDOT’s Materials Division, the LCCA model

developed in this study explicitly recognizes the fact of potential subbase layer deterioration

because of contamination by subgrade fines, comporting with the widely recognized function of

geotextile separator.

PURPOSE AND SCOPE

The purpose of this study was to evaluate the current state of the practice regarding the

use of geosynthetics in road construction and to provide guidance in terms of long-term benefits

to VDOT regarding their use. There was no attempt to attribute structural value to geosynthetics

or to modify pavement design procedures currently employed by VDOT. The main emphasis

was the separation and stabilization functions provided by geosynthetics.

The scope of the study was limited to formulating best practices based on a literature

review; evaluating the long-term performance of two sites with geosynthetics in Virginia;

estimating the potential cost-effectiveness of separator geotextile in secondary pavements by

means of a cost model that conceptually incorporates pavement layer deterioration; and

identifying potential topics for future research. The study was focused on low-volume roads, as

they comprise the vast majority of the road network managed by VDOT.

3

METHODS

Overview

Four tasks were carried out to achieve the study objectives:

1. A literature review was conducted on the current state of the practice with regard to

the use of geosynthetics in road construction. The review focused on peer-reviewed

studies and literature sources and specifications and practice guidelines from various

state departments of transportation (DOTs) and other transportation agencies.

2. Current VDOT state of-the-practice information was gathered from VDOT

specifications and the MOI. In addition, a survey of VDOT resident engineers was

conducted to assess current practices and outcomes on low-volume roadways.

3. Two sites with geosynthetics in Virginia were analyzed using falling weight

deflectometer (FWD) equipment, field borings, and test pits. One site is in Bedford

County, at the junction of Routes 616 and 757. The other site is part of the

realignment of Route 743 in Albemarle County. The sites were constructed in 1994

and 2007, respectively. Field test results were compared with the in-service findings

from previous years.

4. An LCCA method was developed that incorporates pavement layer deterioration

detail that is absent from the LCCA method currently used by VDOT’s Materials

Division. The development of reasonable pavement deterioration curves was an

essential input to the proposed LCCA model. The benefit of geotextile separator was

quantified in terms of reduced pavement subbase deterioration over the analysis

period. As far as practicable, VDOT’s construction database was used to determine

the actual unit costs of construction and maintenance items as functions of the

quantities purchased.

Literature Review

The literature regarding geosynthetics use in pavement structures at the subgrade level

was identified by use of the resources of the VDOT Research Library and the University of

Virginia Library. Online databases searched included Transport Research International

Documentation (TRID), the Engineering Index (EI Compendix), Transport, and WorldCat,

among others. In addition to a search of the websites of the select state DOTs, pavement and/or

geotechnical engineers of several DOTs were contacted through e-mail to determine DOT

practices and the DOT’s specification with regard to geosynthetics use in pavements. The

practice guidelines were synthesized from personal communications with the DOT

pavement/geotechnical engineers, specifications, and the literature.

4

Determination of State of the Practice in Virginia

The information regarding VDOT’s current state of the practice was gathered by

consulting VDOT field engineers and examining the geosynthetics specification in Section 245

of VDOT’s Road and Bridge Specifications (VDOT, 2016). In addition, VDOT’s Pavement

Design Guide (VDOT, 2018a) and the MOI were examined.

A survey of VDOT resident engineers was also conducted to assess current practices and

outcomes for low-volume roadways. Survey questions and responses are presented in Appendix

A.

VDOT specifications and practices were compared to the specifications of other U.S.

highway agencies. Recommendations were made to amend the VDOT specifications to provide

for the effective use of geosynthetics in separation and stabilization applications.

Analysis of Test Sites With Geosynthetics in Virginia

The field performance of two VDOT test sites in which geosynthetics were used at the

interface between the unbound base and subgrade was evaluated. Two documented sections on

VDOT’s secondary road system were used for this evaluation: (1) a section in Bedford County,

and (2) a section in Albemarle County. These sites were selected because they were the only two

known and documented geotextile separator pavement sections on the VDOT network.

The experimental sections in Bedford and Albemarle counties had been in service for

approximately 24 and 10 years, respectively. The Bedford County section is near Smith

Mountain Lake, and the section in Albemarle County is in proximity to the Charlottesville-

Albemarle Airport. The construction details of the two sections were extensively documented in

research reports from Virginia Tech (Appea, 1997; Bhutta, 1998) and the Virginia Transportation

Research Council (VTRC) (Hossain and Schmidt, 2009). The field performance of these

sections was evaluated using pavement condition data, including visual (automated video

logging) distress survey and FWD deflection data. The Bedford County site was also excavated

to expose the base course aggregate to determine contamination from subgrade soil using

gradation analysis and X-ray diffraction (XRD) techniques. The exhumed geotextile was also

tested for retained tensile strength.

Bedford County Site

The Bedford County test sections were constructed by VDOT, as part of a Virginia Tech

research project, in June 1994 on Route 757, at the intersection of Route 616, and included nine

50-ft-long test strips with a base thickness of 4, 6, and 8 in. Sections with each of the three base

thicknesses included a control section (no geosynthetic), a geotextile section, and a section with

geogrid at the interface between the base and subgrade. These sections were documented and

monitored for performance for 8 years by researchers from Virginia Tech. Figure 1 shows the

layout of all nine sections.

5

Figure 1. Bedford County Geosynthetics Test Sections

The predominant subgrade soil type was reddish-brown fat clay (Unified Soil

Classification System [USCS] designation CH, AASHTO A-7-6) except for Sections 4 and 5,

where yellowish-brown silt (USCS designation ML, AASHTO A-5) was observed. The

optimum water content (OWC) and maximum dry density (MDD) for the modified Proctor test

along with other important soils properties, such as liquid limit, plasticity index, percent passing

the No. 200 sieve, specific gravity, and subgrade California bearing ratio (CBR) value, were

determined as follows:

Fat clay (CH): MDD = 100.6 pcf; OWC = 24.4%; liquid limit = 56 to 68; plasticity

index = 28-37; percent passing the No. 200 sieve = 76%; specific gravity = 2.77;

soaked CBR = 7

Silt (ML): MDD = 108.9 pcf; OWC = 17.0%; liquid limit = 41; plasticity index = 4 to

6; percent passing the No. 200 sieve = 73%; specific gravity = 2.74; soaked CBR = 7

Field compaction: water content (WC) = 29% to 36%; dry density = 78.3 to 89.1 pcf.

A local limestone aggregate satisfying VDOT 21B gradation was used as the base course

in all sections with variable thicknesses as follows: 4 in for Sections 1, 2, and 3; 6 in for Sections

4, 5, and 6; and 8 in for Sections 7, 8, and 9. Aggregate properties were OWC = 6.1%; MDD =

142.6 pcf; SG = 2.78; field dry density = 143.2 pcf; and field WC = 2% to 3.2%.

6

A polypropylene-based woven geotextile was placed on the subgrade soil in Sections 2,

5, and 8, and a biaxial geogrid was used in Sections 3, 6, and 9. Approximately 3.5 in of hot-mix

asphalt (HMA) was placed in two layers over a thin layer of chip seal with a 0.5-in-diameter

maximum stone rolled with a vibratory roller. The HMA was a VDOT SM-2A mixture, which is

a Marshall 50 blow mixture with air voids = 3% to 6%; stability = 5500 N; and flow = 2 to 4 in.

The road has an annual average daily traffic (AADT) of approximately 1,150 with 4%

trucks. The traffic count for the last 24 years was collected from a VDOT database and

converted to equivalent single-axle loads (ESALs). The load factors used for this conversion

were 0.0002, 0.46, and 1.05 for car, single unit, and tractor-trailer, respectively, and were typical

for VDOT. There was a difference between the traffic count on Route 616 and Route 757; the

approximate ESALs applied over 24 years were 62,000 and 90,000, respectively. Sections 1

through 3 were on Route 616; Sections 7 through 9 were on Route 757; and Sections 4 through 6

were at the intersection. An additional 10,000 ESALs was applied to these sections as part of the

Virginia Tech study (Al-Qadi and Appea, 2003) over a 2-week period in 1996.

The sections were located on a 4% profile grade with drainage ditches on both sides. No

drainage system was provided. Since this intersection was on a radius, the sections were

constructed with the required superelevation. During the last 24 years, the total rainfall amount

was 1,066 in, but the average annual rainfall was highly variable (coefficient of variance = 50%)

with a mean and standard deviation of 46 in and 23 in, respectively. It is important to note that

rainfall was unusually high in the first 2 years: in 1995, 129 in, and in 1996, 90 in. On the other

hand, the traffic count was very low for both of these years, i.e., below 400 AADT.

The performance of these sections was monitored by the Virginia Tech research team

over the first 8 years using distress ratings (mainly rutting), ground-penetrating radar, FWD

measurements, and other in-place instrumentation.

Albemarle County Site

The second Virginia test section is on Route 743 near the Charlottesville-Albemarle

Airport in Albemarle County, as shown in Figure 2. In August 2007, a 300-ft test section was

constructed with a geotextile at the interface between the base aggregate and subgrade on the

northbound lane of this realignment project. The southbound section, which did not contain

geotextile, served as the control. The construction of this test site was monitored and

documented by VTRC researchers.

There was some variation in the subgrade soil along the length of the pavement. In the

central portion of the test section, the subgrade soil was an elastic silt (MH) with a gradation of

91.7% passing the No. 200 sieve and liquid and plastic limits of 69% and 48%, respectively.

The Proctor MDD and OWC were 80.5 pcf and 37%, respectively. The soil had a soaked CBR

of 4.5. The subgrade soil just south of this test section was identified as sandy clay (CL) or

sandy silt (ML) with an MDD of 97 pcf and an OWC of 17%.

7

Figure 2. Albemarle County Geotextile Test Section

The base course was 8 in of aggregate satisfying VDOT 21B gradation. Another 8 in

of HMA was placed over the base aggregate in multiple layers: 6.5-in BM-25.0 and 1.5-in SM-

9.5A; both were VDOT-designated Superpave mixtures.

A 315-lb tensile strength woven polypropylene AASHTO Class 1 geotextile was placed

at the interface between the base and subgrade in the northbound lane for 300 ft as a test section,

and the corresponding southbound lane was considered a control (no geotextile) section during

the initial construction.

The traffic count on this secondary road is approximately 8,700 AADT, with 2% trucks.

The total applied ESAL count over the last 10 years was 173,000. The annual rainfall in this

area is approximately 40 in, with a total accumulation of 420 in of rain in the last 10 years. One

significant aspect of this test site is the presence of pavement edgedrain (VDOT UD-4 system)

placed along the entire paved section.

8

Development of New LCCA Method

LCCA is a method of transforming the initial and future costs of alternative paths to

accomplish a given outcome into comparable monetary terms in the present. LCCA is

particularly useful when alternatives with the same outcome have variously timed service life

costs over an extended time period. An “analysis period” denoted in years can be a proxy for

service life when alternatives are not distinguished by discrete service lives, as is true for

secondary pavements. LCCA methods are straightforward except for the treatment of “salvage

value.” A complete LCCA should be equipped with an approach for the incorporation of the

value of service life remaining at the conclusion of the analysis period for each alternative under

consideration. Two analytical approaches were evaluated here for incorporation of salvage

value.

Currently, there are no commonly accepted empirical LCCA models designed to evaluate

geotextile separators. There is a degree of consensus that the benefits of geotextiles are self-

evident even if unquantified (Christopher, 2014). For example, when soft soils are unexpectedly

encountered, the cost savings of stabilizing geosynthetics are so widely recognized that

geosynthetics are now considered by some to be the standard of practice (Christopher, 2014).

Similarly, advocates hold that separator geotextile used on a competent subgrade without an a

priori fines problem will result in maintaining long-term pavement performance.

Theoretical long-term benefits of using geotextile separators in pavement structures were

identified nearly 25 years ago. Al-Qadi et al. (1994) determined a traffic benefit ratio (TBR)

offered by pavement with geotextile relative to pavement without geotextile. The TBR consisted

of an increase of 1.7 to more than 2.5 times the design ESALs that a pavement with geotextile

could absorb before 25 mm of rutting damage was attained (Al-Qadi and Appea, 2003). Similar

results and corresponding TBR values were derived by Perkins (2001). Al-Qadi et al. (1994)

concluded that substantial improvement in pavement performance would be possible if

geosynthetics were used for pavement layer separation when the subgrade soil CBR was in the

range of 2 to 6 (Al-Qadi and Appea, 2003).

LCCA was subsequently performed for 25 hypothetical secondary pavements, with

varying traffic levels reflecting secondary roads in Virginia, that were designed in accordance

with 1993 AASHTO pavement design guidelines (Yang and Al-Qadi, 2007). This LCCA

exercise compared agency and user costs for geotextile-free AASHTO pavements with

pavements containing geotextile designed as earlier proposed separately by Al-Qadi and Perkins

for pavement stabilization and reinforcement, respectively. The results indicated that the

incorporation of geotextile for either stabilization or reinforcement purposes could save an

agency up to 40% but might also provide no savings over a geotextile-free pavement. A

significant contribution of the analysis was its approach to translation of theoretical TBR gains

into quantitative cost savings.

Although some empirical case studies reporting long-term savings resulting from the use

of separator geotextiles in unpaved roads have been published (Hawkins, 2009; Whitaker, 2014),

they are not applicable to this study even though good experimental conditions may have existed

because the monetary benefits of geotextiles for paved roads are not simple multiples of the

9

benefits for unpaved roads. However, a 2018 study from the National Center for Asphalt

Technology (NCAT) (Robbins and Tran, 2018) directly influenced the LCCA model developed

in this study. The study reported its own survey finding that most state DOTs performing LCCA

on asphalt and concrete pavements are internally inconsistent between their actual practices and

their LCCA calculations: the majority of agencies used an initial performance period of 10 to 15

years, and the average asphalt pavement age at time of first rehabilitation was approximately 18

years, according to the authors’ evaluation of Long-Term Pavement Performance (LTPP)

Program data.

The survey revealed that “[p]ractices for determining the actual timing of the first

rehabilitation for both [asphalt concrete] and [portland cement concrete] pavements are unique to

each DOT” and that “there does not appear to be a nationwide consensus among DOTs on

[International Roughness Index] values which indicate the need for first rehabilitation” (Robbins

and Tran, 2018). The survey findings supported the decision to adopt a realistic initial

performance period of 20 years in the LCCA model developed in this study for low-volume

secondary roads.

The Pennsylvania DOT (PennDOT) published an example of an economic analysis that

quantified the benefit of geotextile separation in one-third-scale pavement models under

simulated loading, finding a cost savings of at least 13% for a collector road (Petrasic, 2017).

That analysis and the analysis performed in this study used AASHTO (1993) for all pavement

designs; both analyses also assumed that geotextile provides no reinforcing function. However,

this analysis differed from the PennDOT approach in several critical aspects. First, the

PennDOT analysis contrasted the costs of geotextile with the costs of increasing subbase

thicknesses but this analysis did not address increases in pavement thickness as an alternative to

geotextile. Second, whereas the PennDOT analysis assumed geotextile prevented decreasing

subbase strength, this analysis assumed that subgrade strength remains constant while geotextile

prevents decreasing effective subbase thickness. Third, the PennDOT analysis assumed a single

rate of contamination and an initial CBR of 5 whereas this analysis set CBR to 3.3 and tested a

range of contamination rates with the lower bound at PennDOT’s assumed contamination rate.

Fourth, this analysis compared pavements with and without geotextile over a 50-year analysis

period by conventional discounting of future costs to present value whereas the PennDOT

analysis compared pavements with and without geotextile at the conclusion of a 20-year period

of service and inflated future costs by a presumed inflation rate.

The LCCA model developed in this study compared two pavement design options

corresponding to the initial AADT values of 155 and 1,500, each subject to four alternative

scenarios:

1. both separator geotextile and preventive maintenance

2. preventive maintenance but no separator geotextile

3. separator geotextile but no preventive maintenance

4. neither separator geotextile nor preventive maintenance.

Each scenario was analyzed under three different contamination rates. Thus, the entire analysis

encompassed 24 different hypothetical scenarios (2 AADT design x 4 options x 3 contamination

10

rates). Consistent with the observed VDOT practice for rural secondary roads, surface mixture

overlay on the existing road surface was considered typical activity when preventive

maintenance was not provided.

The two pavement design options are consistent with the MOI, Chapter 6 (VDOT

2018b). They were created by means of the 1993 AASHTO design method (AASHTO, 1993) in

conjunction with the values provided in Table 1. The year 2020 was used as the initial design

year (end of construction). AADT and truck % shown in Table 1 for the two pavement designs

cover the typical range of low-volume roadways. The subgrade resilient modulus value is

equivalent to a CBR value of 3.3.

An initial service life of 20 years was selected as realistic, regardless of AADT, for a

Superpave surface mixture on a secondary road, despite the ideal of an initial service life of 12

years. Separator geotextile was assumed to provide benefits consistent with the findings of the

literature review: specifically, it prevents the source of pavement deterioration caused by

aggregate contamination by subgrade fines. The LCCA model develops reasonable maintenance

activities and schedules that respond to deterioration of the four pavement design options over

the analysis period. Slower deterioration because of the presence of geotextile translates into

longer service life after initial construction and longer maintenance cycle spacing for each

subsequent maintenance activity as compared to a pavement without geotextile.

Appendix B describes the development of the pavement layer deterioration curves and

maintenance schedules. Deterioration curves developed for this LCCA model incorporate more

detail on structural deterioration, such as base mixture stripping and permeability loss because of

contaminated base aggregate layers, than do the deterioration curves that forecast network-wide

pavement surface conditions for the Pavement Management System (PMS) administered by

VDOT’s Maintenance Division. Because empirical pavement layer deterioration curves with

and without geotextiles are not currently available, idealized deterioration curves were developed

for this study under the following assumptions:

Table 1. Pavement Design Input Values for the Purposes of LCCA

Pavement Design Input

Values for Low Volume

Pavement Design

Values for Very Low

Volume Pavement Design

Annual average daily traffic (2020) 1,500a 155a

Performance period (yr) 20 20

Growth rate (%) 2.0a 2.0a

Tractor trailers (%) 5.0a 1.0a

Single unit trucks (%) 1.0a 0a

Trucks in design direction/lane (%) 50/100 50/100

ESAL factor Car 0.0002

Single unit truck 0.46

Tractor trailer 1.05

Car 0.0002

Single unit truck 0.46

Tractor trailer 1.05

Initial serviceability 4.0 4.0

Terminal serviceability 2.5 2.5

Reliability (%) 75 75

Overall standard deviation 0.49 0.49

Subgrade resilient modulus (psi) 5,000a 5,000a

LCCA = life cycle cost analysis; ESAL = equivalent single-axle load. a Values assumed for the initial pavement design; all others are in accordance with the VDOT Materials Division’s Manual of

Instructions, Chapter 6, for Farm to Market Secondary Route.

11

Structural coefficients for surface, intermediate, and base layer asphalt mixtures are

based on remaining life factors (see Condition Factor in Figure B2 in Appendix B)

available in the AASHTO Guide for Design of Pavement Structures (AASHTO,

1993).

All HMA mixtures have an initial structural coefficient of 0.44, in accordance with

the MOI (VDOT, 2018b). However, in every pavement alternative analyzed in the

LCCA, surface mixtures degrade over a period of 20 years, as discussed previously

(allowing a 67% margin above the 12 years stipulated in the MOI, Section 607),

whereas base and intermediate mixtures degrade over a period of 30 years (based on

empirical experience at VDOT).

Surface mixture condition over a pavement structure in “poor” condition with

structural coefficients below 0.33 exhibits faster deterioration than over a pavement

structure in better condition.

Base mixture deterioration for an aggregate layer with diminished permeability (i.e.,

with infiltration of fines into the aggregate layer because of the absence of separator

geotextile) is increased relative to a pavement with geotextile in order to replicate

pore water increase and potential stripping.

No structural benefit is attributed to the presence of separator geotextile.

Proper characterization of the separator geotextile benefit was essential for the

determination of pavement maintenance schedules. In this study, the benefit of geotextile was

modeled as delayed aggregate contamination (or, the equivalent, delayed loss of subbase

functionality). Pavement layer deterioration curves presented in Appendix B and used in the

LCCA model are based on AASHTO deterioration curves but are also differentiated by the rate

of subbase deterioration in the form of effective aggregate thickness reduction. The rate of

contamination is therefore an exogenous parameter of the model that influences the timing of

maintenance schedules and the type of activities required in pavements without the geotextile

separator. This mechanism comports with the study findings on the beneficial function of

separator geotextile and allows clarity in pavement design calculations. A range of aggregate

contamination rates in pavements without geotextile was tested in the LCCA model to span a

realistic range of field conditions.

Generally, maintenance decisions are triggered when the structural coefficient drops

below 0.30 on the surface mixture deterioration curve, thus indicating “poor” surface condition.

Patching activity is triggered when the remaining design life falls below 5 years but the surface

structural coefficient is still above 0.30 (based on 1993 AASHTO pavement design) (AASHTO,

1993). The LCCA model developed in this study is conceptually similar to the pavement-related

LCCA guidance provided at the national level. In addition, the Federal Highway

Administration’s (FHWA’s) RealCost Version 2.5 software requires users to supply pavement

service life estimates at construction and for every subsequent activity to be included in the

LCCA.

12

Appendix C provides an example of the decision process for determining the

maintenance schedule for one of the four pavement options analyzed in the LCCA. Appendix D

contains the detailed maintenance schedule of activities and costs over the 50-year analysis

period for the same pavement option, as used in the LCCA.

Final LCCA results consist of the sum of construction and maintenance costs at the end

of the analysis period discounted to their present value plus a factor that considers the varying

remaining service lives of the pavements at 50 years. Actual VDOT data were used to determine

construction and maintenance costs for each pavement design, including the pavement design

that incorporates geotextile. Because documentation of the use of geotextiles for separation is

not currently required by VDOT, identification of actual costs involved two subordinate tasks:

first, the choice of an item code for the geotextile typically used for separation purposes in

VDOT roads; and second, selection of a subset of construction projects in which the separator

GTX was used on the secondary system. When the tasks were completed, quantity-to-unit price

relationships were identified for the geotextile and for most major items in secondary pavement

construction. These costs were also used as representative for maintenance activities.

A discount rate of 4% was used as the base case, consistent with the MOI, Chapter 6

(VDOT, 2018). The life cycle cost calculation for a pavement option, assuming that interest was

compounded annually, is summarized in Equation 1.

LCC = (∑Costi

(1+d)ii=Ti=0 ) ± 𝑓 [Eq. 1]

where

i = years from project start Costi = project cost in year i T = analysis period in years

d = discount rate

𝑓 = remaining service life factor at 50 years.

Judicious accounting for the remaining service life factor is an aspect of LCCA about

which there is no clear consensus in practice. However, pavement alternatives will almost

certainly exhibit differing states of adequacy (surface asphalt condition, structural condition,

estimated remaining design life, etc.) at the end of an analysis period. To incorporate better or

worse final conditions into the present value of the cost of each alternative, some kind of

normalizing of different final states must be performed. In this study, two methods are offered in

which remaining design life is treated as either (1) an asset (salvage value) or (2) a “normalized”

liability reflecting differing final conditions.

In the first method, salvage value is calculated for each alternative pavement design as

the product of estimated remaining design life at the beginning of year 50 (prorated over 20

years) and the cost of a 20-year pavement design (constant for all alternatives) given the AADT

in year 50 and discounted to the present. Salvage value is monetized by this method and is

treated as a life cycle cost offset; i.e., the salvage value factor is subtracted from the sum of

construction and discounted maintenance costs up to year 50.

13

In the second method, the modeled condition of each existing pavement structure after 50

years is the basis for the creation of a 20-year “terminal pavement design” for each pavement.

Pavements with a low estimated remaining design life require more rehabilitation, at probably

higher total cost, to achieve a 20-year design life than pavements with a higher estimated

remaining design life at year 50. The costs of a terminal design for each pavement after year 50

are discounted in the same manner as earlier maintenance activity costs and added as another

liability, of greater or lesser magnitude depending on condition at year 50, to the sum of

discounted construction and maintenance costs through year 50.

Although the methods vary conceptually and lead to different magnitudes for life cycle

costs, the relative results are consistent between methods. Figure 3 summarizes the LCCA

process in general terms.

Last, the discount rate is sometimes assumed to be an exogenous input that can be set

near 3% with little supporting argument or repercussion even though it plays a critical role in

LCCA results. In fact, the discount rate at its simplest is a calculation of the expected nominal

interest rate less the expected inflation rate over the analysis period. Thus, to be meaningful, this

figure is tied to current and expected future economic conditions. In the simplest interpretation,

the discount rate should represent the “real” or inflation-adjusted earnings that are possible if

project funds were invested rather than spent on the project. Yet explicit price inflation over the

analysis period, at best a speculative undertaking to predict for the future, would be challenging

to forecast in detail for LCCA calculations.

To solve this issue consistently, the choice of the LCCA discount rate to be used by a

state highway agency is typically guided by FHWA recommendations. For LCCA of pavement

designs, FHWA favors a discount rate that represents a long period of time such as historical

values reported periodically in Appendix A of Circular A-94 issued by the U.S. Office of

Management and Budget (OMB) (FHWA, 1998). In 1992, the OMB revised policy by requiring

that applied discount rates be updated when the interest and inflation rate assumptions used in

preparation of the federal budget are changed (Federal Register, 2018). In other words, the

choice of a discount rate should be consistent with FHWA recommendations and therefore with

OMB guidance for cost-effectiveness analysis of federal programs in a given fiscal year, rather

than assumed to be a particular fixed value.

In November 2018, the OMB recommended an applied real discount rate of 1.5% for a

30-year analysis period (OMB, 2018), in contrast to a few years earlier when the required

discount rate for federal transportation grant applications was 7%. Reflecting stated OMB

policy, the recent figure of 1.5% represented real interest rates on 30-year treasury notes and

bonds of the maximum available specified maturities as incorporated in the federal budget for

2020. Two discount rates were used in this LCCA: 4% was used to place the rate in the range of

the existing LCCA of VDOT’s Materials Division, and 1.5% was used to comply with the OMB

recommendation. The application of a discount rate that is lower than the base case, however,

will often preserve the relative LCCA outcomes resulting from the higher rate.

14

Figure 3. Outline of Life Cycle Cost Analysis Method. Assumed constants = resilient modulus, performance period, initial annual average daily traffic

(AADT) and growth rate by scenario, equivalent single-axle load (ESAL) factors by vehicle type, initial and terminal serviceability, reliability, truck%

of AADT by scenario, drainage coefficient, and standard deviation.

15

RESULTS

Literature Review

There are four types of geosynthetics commonly used in roadway design (Equipment

World, 2014):

1. Geotextiles are either woven or nonwoven fabrics. They are used for separation,

filtration, stabilization, and reinforcement. A subset of geotextiles are paving fabrics

used between pavement layers. Woven fabrics are typically used in reinforcement

applications. Nonwoven geotextiles are used mainly for separation and filtration. In

practice, separation, filtration, and stabilization functions provide composite action,

although depending on the site conditions one is typically more dominant than the

others. The stabilization or reinforcement function of a geotextile stems from

providing lateral restraint of the base and subgrade. Geotextile allows for interface

friction forces to develop, thus reducing horizontal deformations and increasing

stability.

2. Geogrids are mesh-like materials formed by a network of tensile elements with

openings of sufficient size to allow interlock with the surrounding granular fill

materials. Geogrids are primarily used for reinforcement and confinement. They can

provide reinforcement in one direction (uniaxial), in two directions (biaxial), or in

three directions (triaxial). The performance of geogrid as reinforcement depends on

its rigidity and geometry of openings, which controls the interlock with soil and base

materials.

3. Geocomposites are a combination of two or more geosynthetic materials, such as

geotextiles with a core, used to enhance drainage, for example, as prefabricated

longitudinal edgedrains or a layer feeding into these drains.

4. Geomembranes are a single, solid sheet of geosynthetic material, used in construction

as an impermeable barrier.

In the United States, the first reported use of a cotton textile in road subgrade stabilization

was by the South Carolina Highway Department in 1926 (Beckham and Mills, 1935), but a

widespread use of synthetic textiles coincided with the developments in Europe in the 1970s.

The main objective was to prevent the intermixing of two adjacent dissimilar materials, so that

the integrity of the materials on both sides of the geotextile remained intact. Vosteen and

Wilmers described the extensive use of nonwoven geotextiles in the construction of the A-45

motorway in Germany in the mid-1970s (cited in Raymond and Giroud, 1993). Geotextiles were

used as separators on soft subgrade soils. Successful results were reported. King reported on the

use of separator geotextile in the construction of the M-11 motorway in the United Kingdom in

1978 (cited in Raymond and Giroud, 1993). The fabric was placed on a clay/silt subgrade with

low CBR values. Subsequently, data gained from this project served to develop Clause 609

specifications for the use of geotextiles as separators in the United Kingdom.

16

According to Koerner (2005), the main advantages of geosynthetics are as follows:

They are quality-control manufactured in a factory environment.

They can be installed rapidly in many cases.

They generally replace natural resources.

They generally replace different designs using soil or other construction materials.

They are generally cost competitive versus soils or other construction materials

that they replace or augment.

Their technical database (both design and testing) is reasonably established.

Geosynthetics are manufactured from polymeric materials that mainly include polyester,

polypropylene, polyethylene, polyamide, and polyvinyl alcohol (Greenwood et al., 2012). No

product is made of 100% resin. The primary resin is typically mixed with chemical additives

serving as antioxidants, thermal stabilizers, and ultraviolet light absorbers. The total amount of

each additive varies widely depending on the formulation developed by a manufacturer.

Geosynthetics in all applications should be tested for resistance to weathering. Although

geosynthetics contain carbon black to increase their ultraviolet resistance, they are considered

unsuitable for permanent exposure to sunlight. The assessment of durability of a geosynthetic is

often made with certain reduction factors applied to key properties used in design. In general,

the subject of durability of geosynthetic materials used on various construction projects has not

been reported as a critical issue to date, with most failures traced to the incorrect material

selection or faulty installation. Geosynthetics have been shown to be essentially inert materials

for most environments and applications.

The tensile strength of most geosynthetics is generally not the same in all directions, as

they behave as anisotropic materials (Shukla, 2016). In woven geotextiles, tensile strength is

governed by the weaving structure. The warp strength (machine direction) is usually greater

than the weft strength (cross-machine direction). Typically, woven geotextiles display lower

extensibility and higher strength as compared with nonwoven geotextiles. Geogrids normally

exhibit high dimensional stability, high tensile strength, and a high tensile modulus even at low

strains, which allows them to be effective in reinforcing applications. Many geosynthetics have

a viscoelastic stress-strain behavior. It is generally found that the modulus of a geosynthetic

confined in soil is likely to be higher than when tested without a confining load.

The commercial availability of geotextiles allows for economical construction on

marginal soils. Geosynthetics produced in a factory environment have specific and uniform

properties that can address challenging soil conditions. With a soft subgrade, it is often

impossible to build a stable base course without losing expensive base material into the subgrade

(Van Santvoort, 1994). Geosynthetic layers are placed during the initial construction phase,

typically at the interface between the subgrade and base course and sometimes within the base

course. Frequently, the primary objective is to allow access for heavy construction equipment

17

and to provide a stable platform to allow compaction of granular layers. Significant cost savings

can be realized in the early stages of road construction by minimizing the loss and waste of

aggregate base in the soft subgrade soil. The development of various geosynthetic materials

provided new options for constructing and enhancing performance.

There are two prevalent approaches to geosynthetic design, typically addressing either

function or specification. Design by function is based on the calculation of the required material

property with an added factor of safety (Koerner, 2016). In design by specification, several

application categories are listed in association with various physical, mechanical, hydraulic,

and/or endurance properties. The resulting material designation is determined by the intended

application. This method was originally advanced by an AASHTO task group and resulted in the

AASHTO M 288 geotextile specifications (Suits and Richardson, 1998). Government agencies

generally follow AASHTO M 288, often incorporating local modifications. Design by material

specification is based on decades of experience gained under various service conditions.

AASHTO M 288 covers six geotextile applications: subsurface drainage, separation,

stabilization, permanent erosion control, sediment control, and paving fabrics. However,

AASHTO M 288 is not a design guideline. It is the engineer's responsibility to select a

geotextile for the application that takes site-specific soil and water conditions into consideration.

AASHTO M 288 defines three classes of material strength. Class 1 designates severe or

harsh survivability conditions. Class 2 designates typical conditions and is used as a default

classification. Class 3 designates applications where there is little or no potential for

geosynthetic damage. Strength is considered the principal property required for a geosynthetic to

survive the installation and provide the required functionality.

AASHTO M 288 stipulates the use of geotextile separator for pavement sections

constructed over unsaturated subgrade soils with a CBR value greater than 3. It provides the

corresponding material and strength class properties. Of significance, AASHTO M 288 does not

set the upper CBR limit on the applicability of geotextile separators. When the subgrade soils

are saturated and CBR values range between 1 and 3, stabilization becomes the primary function.

AASHTO M 288 provides the required geotextile requirements for pavement structures

constructed in these conditions. The specification is not appropriate for embankment

reinforcement, where global stability is of concern. In addition, very soft subgrades with CBR

values below 1 are not addressed by AASHTO M 288, and geotextiles are not recommended in

such cases. AASHTO M 288 permittivity limits on separation and stabilization applications are

0.02 sec-1 and 0.05 sec-1, respectively.

AASHTO M 288 allows the use of woven slit-film geotextiles in separation and

stabilization applications but prohibits their use in subsurface drainage and erosion control.

These fabrics are produced by weaving individual flat yarns that are cut from an extruded film of

synthetic material. Their permittivity and surface friction are relatively low when compared to

those of other woven fabrics. The U.S. Bureau of Reclamation does not recommend the use of

slit-film geotextiles in embankment dams (Gobla, 2014). Unified Facilities Criteria (UFC) for

engineering use of geotextiles state that slit-film geotextiles are not preferred because their

opening sizes are unpredictable (U.S. Department of Defense, 2004). Long-term clogging is a

18

concern when these geotextiles are used for filtration. The Montana DOT (2008) is one of the

transportation agencies that explicitly prohibits their use for subgrade stabilization.

Testing of geosynthetic products used by DOTs is commonly conducted through the

National Transportation Product Evaluation Program (NTPEP) under the auspices of AASHTO

(AASHTO, 2018). In addition, the program provides on-site inspections of plants providing

products manufactured to AASHTO standards. Currently, there is an ongoing NTPEP task force

effort aimed at developing material specifications for the use of geogrids and geotextiles in

subgrade stabilization, with the possible inclusion in the future AASHTO M 288 standard.

The FHWA published a comprehensive set of design and construction guidelines for

geosynthetics by Holtz et al. (2008). The guidelines are directed to highway engineers as a

means to select and specify geotextiles and geogrids properly. Stabilization is defined as the

primary function when the subgrade CBR value is less than 3. Separation and filtration are the

primary functions for CBR values ranging between 3 and 7. Additional guidance on the use of

geosynthetics in pavement construction dealing with the geotechnical aspects of pavements is

provided in Christopher et al. (2006).

There are numerous reports indicating that geosynthetics provide a strengthening effect.

Various authors have reported a potential reduction of the design base thickness while the

original design load carrying capacity of pavements is maintained (Shukla, 2016). Rankilor

(1981) recommended that this should not be the key argument for using a geotextile separator.

In his opinion, the main economic benefit is derived from savings on the aggregate base material

that would otherwise be lost in the soft subgrade during construction.

Van Santvoort (1994) reported that the upper threshold for using geotextile should be 1

tsf undrained shear strength at the maximum water content. He referenced practical experience

from The Netherlands indicating that above 1 tsf no particles in the subgrade are loosened and

pumped into the base course.

A common argument is that the presence of a geosynthetic separator layer decreases the

potential for base contamination by fines migrating from the underlying soft soil subgrade. As a

consequence, contamination effectively leads to a reduced base layer thickness and, ultimately,

to a decreased road life (Zornberg, 2017). Al-Qadi et al. (1994) stated that in the absence of

separation, two simultaneous mechanisms tend to occur over time in pavements. Soil fines

migrate into the base course voids, resulting in a diminished drainage performance, and the base

course particles penetrate into the subgrade, compromising the pavement structure. The resulting

reduction in the effective thickness of the aggregate base was studied extensively by Jorenby and

Hicks (1986). Their laboratory study concluded that up to 6% added fines can be tolerated

without markedly affecting the base material stiffness and that the primary benefit derived from a

geotextile is the increased service life of pavement structure. Jorenby and Hicks also reported

that 8% fines was sufficient to disrupt the functioning of an aggregate drainage layer.

Koerner (2005) stated that separation is the most underrated of all geotextile functions

because although every field use involves separation, rarely is separation designed on its own

merit. When a geotextile is used as a separator, it is placed over the soft subgrade. It acts as a

19

filter to allow water but not fine particles to flow through it, preventing any mixing of the soft

soil and granular base under the action of the construction equipment and subsequent traffic.

When separation is considered the primary function in a paved road, an adequate survivability

criterion is required to ensure durability in service.

A recent laboratory study by Kermani et al. (2018) reported an approximate 30%

reduction in the amount of pavement rutting when geotextile separator is used on top of subgrade

under controlled drainage conditions. An accelerated pavement testing device (the MMLS3) was

used to simulate the cyclic traffic loading on a scaled model of a flexible pavement. The

laboratory study concluded that geotextile significantly reduces pumping of subgrade fines into

the granular subbase material. At the end of the test, the percentage of fines that migrated to the

subbase, based on the percentage mass of subbase, was 6.39% in the tests without geotextile and

1.81% in the test containing geotextile. Kermani et al. (2018) also documented how quickly the

contamination process progresses. Pennsylvania No. 2A dense-graded subbase material was

used in the study. Comparable results can be expected in Virginia as VDOT 21B dense-graded

aggregate has a similar design range of particle gradations.

The phenomenon of fine-grained subgrade infiltration caused by traffic-induced cyclic

stresses was studied extensively by Woods and Shelburne (1943) in their pioneering work on

pumping of jointed concrete pavements. Severe pumping at joints and cracks was observed

during periods of heavy rainfall. The upkeep of pumping pavements was considered a serious

challenge confronting U.S. highway engineers during the wartime period. The problem was

further compounded on roads carrying large volumes of heavy traffic, commonly associated with

military transports. Positive correlations were traced to plastic soils, primarily clays and silty

clays. The predominance of pumping was discovered in the cut sections with saturated

subgrades. The problem was largely confined to the heavily traveled concrete roads. At the

time, the recommended solutions included improvements to the subsurface drainage and the use

of properly graded aggregate to prevent clogging. Currently, this problem can be effectively

managed with the use of geosynthetics. For example, in Germany, nonwoven geotextile

separators are routinely placed under all new concrete pavements (Sarsby, 2006).

Geotextiles are used extensively in the railway industry. In most cases they operate in a

significantly harsher environment than most highway geotextiles. They are commonly installed

under the sub-ballast layer to maintain the integrity of railway sections by preventing the

pumping of subgrade fines and the resulting sub-ballast contamination. The railway industry,

recognizing the commonality of the pumping problem, consolidated the findings of Woods and

Shelburne with their historical field observations and identified pumping-susceptible soils as

shown in the plasticity chart in Figure 4 (American Railway Engineering Association, 1946).

The chart axes are plasticity index and liquid limit. Figure 4 depicts a synthesis of documented

case histories where extensive pumping of subgrade fines into the overlying granular layer has

been observed. A point on the chart represents unique plasticity characteristics of a particular

soil. It can be seen that field data are clustered parallel to the Casagrande A-line distinguishing

clays from silts.

20

Figure 4. Plasticity Chart for Pumping Soils Beneath Highway Pavements and Railways. Source: American

Railway Engineering Association, 1946.

Duong et al. (2014) further investigated the pumping phenomenon in railway

substructure. Significant migration of fine particles was observed only under near-saturated

subgrade conditions. On the contrary, under unsaturated conditions, no significant intermixing at

the interface was detected. Currently, the American Railway Engineering and Maintenance-of-

Way Association (2018) recommends the use of nonwoven geotextile separators at the subgrade

and sub-ballast interface. Applications are classified as regular, heavy, and extra heavy, with the

resulting specification defined by weight of fabric of 10 to 12 oz, 12 to 16 oz, and 16 to 20 oz,

respectively.

Black and Holtz (1999) reported on the performance of geotextile separators 5 years after

construction. Excavations were made in the test section of a state highway in Bucoda,

Washington, on a site with a documented history of poor pavement performance. Prior to

placement of the geotextile fabrics, subgrade undercuts of 18 and 24 in were performed on the

northbound and southbound lanes, respectively. The base course was compacted with a smooth

drum non-vibratory roller. The design base thickness was 12 in in the northbound lane and 18 in

in the southbound lane. The asphalt concrete pavement thickness was 6 in. Although some

subgrade fines had migrated through the geotextiles into the bottom of the base course, there was

no evidence of any adverse effect on the pavement performance after 5 years of service. The

researchers concluded that the long-term function of geotextile separators may not be as critical

in many cases because of the increased subgrade strength and reduced compressibility because of

consolidation. Of significance, they observed that subgrade sections beneath geotextiles became

more consolidated than areas without the geotextile. It appears that separation and drainage

functions of geotextiles have significantly more impact on pavement performance in the early

21

stages of pavement life because of consolidation and subsequent strength gain of subgrade soils.

All geosynthetics performed adequately as separators.

In an effort to quantify further the contribution of geotextile separators to the long-term

performance of pavements, the Bucoda test site was investigated at 12 years following

construction (Collins and Holtz, 2005). Extensive field and laboratory tests were conducted on

the subgrade, granular base, and geotextiles. The results of FWD tests revealed that the most

significant increase in the subgrade moduli occurred only in the first few months following

construction. Some of the grain size distribution test results indicated that the fines content in

the base course immediately above the geotextile increased by up to 5%, as compared with the

gradation at higher elevations, but in the opinion of Collins and Holtz (2005), this relatively

small percentage was negligible and would not be detrimental to pavement performance. The

authors concluded that geotextiles may not have provided any significant contribution to the

long-term performance of the pavement section over the 12-year period. This outcome was

likely influenced by the depth of placement (18 to 24 in of subgrade undercut).

Barksdale et al. (1989) studied the potential benefits of geosynthetics in flexible

pavements. Both woven and nonwoven geotextiles were found to perform the separation

function adequately, although thick nonwovens performed better than wovens. Long-term

durability was identified as being of primary importance in separation and filtration applications.

Separation problems are usually most pronounced during construction of the first lift of the

granular layer, when the vertical stress applied to the subgrade is greatest. Large, angular, open-

graded aggregates placed directly on a soft subgrade are most critical. Typically, the total

thickness of the intermixed zone is up to 2 times the diameter of the aggregate material that

overlies the subgrade. The tendency for aggregate penetration greatly increases as the subgrade

water content increases above the optimum value. For a separation problem to develop, the

applied stress level must approach the ultimate bearing capacity of the subgrade, as calculated

with Equation 2:

qult = 5.2 c [Eq. 2]

where

qult = subgrade ultimate bearing capacity

c = subgrade undrained shear strength.

The problem is further complicated by local yielding of soil at lower initial stresses.

Barksdale et al. (1989) recommended that a minimum safety factor of 1.5 to 2.0 be applied to

Equation 2 when no geotextile is used. At one test site where geosynthetics were not used, the

intermixing of subgrade and aggregate base was detected, with the calculated safety factor

ranging between 0.8 and 1.4. The use of separator geotextile to prevent contamination of the

granular layer is recommended in cases where the safety factor is less than 1.4. The practical

implication of these findings is that size, gradation, angularity of particles, subgrade strength,

and the applied stress level are the most significant parameters to consider in selecting a

separator geotextile. In general, subgrades under relatively thin pavement sections, typically

22

associated with low-volume roads, are more prone to experience pumping conditions than

subgrades under heavy structural pavements.

Metcalfe et al. (1995) evaluated the long-term separation and drainage performance of 22

geotextiles. The study focused on various mechanisms adversely affecting geotextile operation,

such as blocking (blinding) on the lower surface and lodging in (clogging) pore openings by fine

particles. In addition, the term “caking” is used to describe the blockage of pore openings,

including chemical precipitation, on the upper surface of geotextile. The authors reported that

some woven slit-film fabrics had significant blinding because of fines migration. Overall, field

observations and laboratory test results indicated that needle-punched nonwovens exhibit the

best long-term drainage performance, probably because of their three-dimensional structure.

Saathoff (1988) evaluated the long-term filtration performance of geotextiles used in

coastal, railway, and highway applications. The geotextiles consisted of 52 nonwoven and 29

woven fabrics that had been in service between 1 and 14 years. The nonwoven geotextiles were

found to have significantly better filtration performance as compared to the woven fabrics.

The use of nonwoven separator geotextile is recommended for permeable pavements at

airports (Bruinsma et al., 2017). Geotextile fabrics are specified to prevent the migration of fines

from the subgrade and contamination of the base layer. An apparent opening size equivalent to

the No. 70 sieve is required for the fabric.

The Virginia Department of Environmental Quality (2013) stormwater design

specifications for permeable pavement include measures to protect the bottom of the reservoir

level from intrusion by the underlying soils. One of the recommended methods to prevent soil

contamination is to use filter fabric separator. The fabric should have an apparent opening size

equivalent to the No. 70 or No. 80 sieve.

Bearden and Labuz (1998) presented results of large scale direct shear tests on soil-

fabric-aggregate systems. Lightweight silt-film woven, heavyweight woven, and heavyweight

nonwoven materials were tested. Test results indicated that nonwoven geotextiles generally had

interface friction angles approximately 20% higher than those of woven fabrics.

Mechanistically, when placed at the top of subgrade in a paved road, the principal

operating mode of a geosynthetic is the lateral confinement effect as opposed to the tension

membrane effect that is commonly associated with unpaved roads with substantial rutting.

Lateral displacement of base aggregate particles, resulting from repetitive traffic loading,

degrades the mechanical properties of base material (Zornberg, 2017). Such displacement is of

particular significance in the lower portion of the base layer, where tensile stresses are prone to

develop. It may be argued that the effective lateral confinement provided by geosynthetics can

reduce aggregate particle degradation or breakage, thus maintaining the performance of the

pavement section over the entire life cycle. The key factor in mobilizing lateral confinement is

interface friction. Nonwoven geotextiles are particularly well suited for this task.

Marginal subgrade can be stabilized chemically, typically using cement, lime, or asphalt

emulsion. The alternative approach is to excavate unsuitable soils and replace them with a

23

competent material. Sometimes, local groundwater conditions may adversely affect the

properties of the backfill material. The procedures of the Nebraska Department of Roads (2012)

stipulate that if high water table conditions exist at the project site, a geotextile is needed to

prevent the soft subgrade material from migrating into the granular layer. Another option

involves the use of open-graded aggregate wrapped with a geotextile to serve as a drainage

blanket below the pavement section. This concept can be implemented in various ways. For

example, the Australian specifications call for blending aggregate sizes of 20, 14, and 10 mm in

diameter with coarse, washed sand (Autostrads Ltd., 2012). This free-draining material can also

be cement stabilized to improve its strength when wet.

Kroening et al. (2017a, 2017b) conducted an extensive study for the Arizona DOT to

develop geosynthetic specifications and application guidelines. The study included a

comprehensive survey of DOTs. Of the 32 responding DOTs, 12 (including VDOT) reported

that they seldom use geotextile fabrics together with geogrids for separation. The most common

combined use of geotextiles and geogrids was for stabilization of soft subgrade soils. The

authors proposed that a 35% fines content, instead of 50%, should be used as a breakpoint

between coarse-grained and fine-grained subgrades. If the subgrade is classified as coarse

grained, the use of separation or stabilization fabric is not necessary. The authors indicated that

currently there is no consensus among the states on the use of AASHTO M 288. Their

recommendations included updating the Arizona DOT geotextile specifications in accordance

with AASHTO M 288, adding a specification for stabilization geotextile, and updating the

Arizona DOT geogrid specifications. Specifically, the authors recommended the removal of the

geotextile thickness because this requirement is already defined by the strength and durability

criteria in AASHTO M 288. A permittivity requirement compatible with Arizona soils was

proposed. The study concluded that although some research suggested that a pavement structure

with geogrid and geotextile will last longer and perform better over time, adequate information is

not currently available to quantify these benefits.

The Minnesota DOT performed extensive field studies on the use of geosynthetics on

county roads in the early 1990s (L. Tasa, personal communication). The research, spanning

more than 10 years, was initiated after several counties reported measurable gains from the use

of fabrics and geogrids. These benefits were identified primarily as the reduction of soft spots

caused by contractors’ equipment and a significantly stronger roadbed during paving operations.

Field observations also included a reduction in the transverse cracking of pavements. The

practice of using geogrids and geotextiles for stabilization and separation has been in effect to

date and the prevalent positive feedback from the counties indicates that the benefits are tangible.

Currently, local projects have been given approval to recognize geosynthetic strength in

pavement design, allowing for the reduction in base thickness.

The Minnesota DOT lists geotextile properties in terms of tensile strength and water

conductivity (Skok et al., 2003). Geogrid materials are specified by tensile strength, junction

strength, flexural stiffness, and torsional stiffness. Special considerations include ductility and

brittleness. Samples of the proposed geogrid must be submitted for approval. Minnesota DOT

Specification 3733 does not distinguish between subgrade separation and stabilization

(Minnesota DOT, 2005). The same geotextile type (AASHTO Class 1) is specified for a

combined use in both applications and listed as a separator geotextile.

24

The Missouri Highways and Transportation Commission (2018) also combines

separation and stabilization functions under the separation label. Specification 1011 for

geotextiles calls for AASHTO Class 1 material with the minimum permittivity of 1.0 sec-1.

The New York State DOT (NYSDOT) developed comprehensive guidelines for

geosynthetic design in their Geotechnical Design Manual (NYSDOT, 2018a). Generally,

NYSDOT follows the guidelines by Holtz et al. (2008) published by FHWA. Design

requirements for various applications are presented, including the separation and stabilization

functions under pavement structures. The manual is explicit in stating that it is not the practice

of NYSDOT to consider the reinforcing effect of the geotextile in pavement design. Geotextile,

if used, is assumed to act purely as a separator. Specifications for geosynthetics, according to

field applications, are compiled in Section 737 of the NYSDOT Standard Specifications.

Geosynthetics considered suitable for specific applications are placed on the list of approved

products (NYSDOT, 2018b).

The Colorado DOT (2005) geosynthetic specifications stipulate that all materials must

conform to the requirements of AASHTO M 288. NYSDOT-approved products are referenced

as a listing of materials that meet the requirements. Separation and stabilization functions are

combined.

The Washington State DOT (WSDOT) design manual addresses separation and

stabilization functions in roadway construction (WSDOT, 2016). Woven and nonwoven

geotextiles are permitted. In general, separator geotextile is specified when the subgrade

resilient modulus values range between 5,800 and 15,000 psi. These values correspond to the

CBR range of approximately 3 to 10 (Heukelom and Klomp, 1962). WSDOT follows the

requirements of AASHTO M 288, Class 2, for separators. Separator geotextile should be used

only on subgrades that can be prepared and compacted as required in the WSDOT Standard

Specifications. There is no need to use separators in dense and granular soils. When the

removal of subgrade soils and replacement with a granular material is required, no separation

fabric is applicable because it defeats the purpose of the geotextile separator. In this case, soil

stabilization geotextile is considered. The stabilization geotextile is placed directly on the soft

subgrade, even if some excavation is required. Backfill material used to replace the over-

excavated subgrade should not be placed below the geotextile. The need for stabilization

geotextile usually arises when the subgrade resilient modulus is less than 5,800 psi or clayey

subgrade is likely to be present. The fabric may also be required if the subgrade is not

anticipated to be saturated based on the available data but may become saturated during

construction. Soil stabilization geotextile should not be used under roadway fills greater than 5 ft

high or on extremely soft subgrades. A site-specific reinforcement design is required in such

cases. WSDOT follows the requirements of AASHTO M 288, Class 1, for stabilization

geotextiles.

The Oregon DOT (2011) Pavement Design Guide refers to Specification 00350,

applicable for all geosynthetics used in construction. The geotextile and the level of certification

must be included in the pavement design recommendations. Typically, nonwoven geotextiles are

used on subgrades with an increased potential for contamination, such as in areas with a high

25

groundwater table. A minimum of 6 in of aggregate material is typically required over the

geotextile.

The Oklahoma DOT pavement design practice includes the universal use of woven

geotextiles as separators (J. Randell, personal communication). Geotextile separators are always

used with granular pavement base to prevent the subgrade fines from intermixing with the base

material. Woven geotextile type is selected depending on the subgrade CBR value, with heavier

fabrics used when low CBR values are encountered. Potential reinforcing properties of the

geotextile are ignored in pavement design calculations.

PennDOT specifies geosynthetics for separation, stabilization, and reinforcement

(PennDOT, 2016). Strength requirements for the separator geotextile exceed those for AASHTO

M 288, Class 1. All separators are relatively heavy 12 oz/yd² nonwoven fabrics, selected to

ensure durability. Potential damage during construction is an important concern for PennDOT.

Permittivity requirements are significantly higher than those in AASHTO M 288 but in line with

the erosion control applications. PennDOT combines separation and erosion control functions

for geotextiles. Separation fabrics are currently used beneath all bituminous and concrete

pavements to prevent fines from migrating into the base (D. Clark, personal communication).

High strength woven geotextiles are used for stabilization. Biaxial geogrids and high strength

wovens are used for subgrade stabilization and reinforcement. Recent modifications to

PennDOT’s geotextile separator specifications were implemented following the study by Xiao et

al. (2016) on the migration of subgrade fines into subbase. The study involved laboratory testing

using the MMLS3 device to simulate migration of fines in the pavement system. The findings

indicated that using a geotextile is very effective in reducing migrations of subgrade soil into

subbase.

In a recent report for NCHRP Project 01-50 (Luo et al., 2017), a methodology was

proposed to quantify the performance of a pavement structure when geosynthetics are used in the

unbound aggregate layer. The primary focus of this study was stabilization and/or reinforcement

benefit in both flexible and rigid pavements. Analytical models developed to quantify the impact

of geosynthetics on pavement performance included (1) a finite element model for computing the

critical stresses and strains, and (2) an artificial neural network model to predict the pavement

performance. Field validation was limited to only a few projects within the LTPP database. The

authors found that the primary advantage of geosynthetic reinforcement was the reduction in the

vertical compressive strain in the base course and at the top of the subgrade. In general, geogrid

was observed to be more effective than geotextile at reducing permanent deformation, but the

reduction was not significant until the deviator stress exceeded 19 psi. No apparent benefit was

detected with the rigid pavement. The long-term influence of subgrade saturation was not

explicitly studied.

Summary of DOT Practices

DOT practices vary but generally follow AASHTO M 288, often augmented with local

modifications. The majority of DOTs developed distinct provisions for separation and

stabilization applications in pavement construction, typically specifying the minimum required

permittivity and the maximum allowable opening size. Class 2 geotextile is typically specified

26

for separation. Class 1 geotextile is prevalent in subgrade stabilization. Woven and nonwoven

fabrics are allowed in both functions, with nonwovens more common in separation.

The optimal breakpoint between separation and stabilization application corresponds to

the CBR value of 3, as stipulated in AASHTO M 288. Koerner (2016) stated that in medium to

firm soil subgrades with soaked CBR values exceeding 3, geotextile functions uniquely as a

separator. There is no clear consensus on the exact upper CBR limit for using separator

geotextile. AASHTO M 288 is silent on this point. DOT practices vary in the CBR range of

approximately 6.5 to 10.

Subgrade stabilization fabrics are typically used on soils with CBR values below 3 and

above 1. Site-specific reinforcement design is required when the CBR drops below 1. Typically,

geotextiles are not assigned any structural contribution in pavement design.

Zornberg and Thompson (2012) conducted a survey of state DOT specifications on the

use of geotextiles in pavements. Table 2 summarizes geotextile specifications based on

application. Table 2. Summary of State DOT Geotextile Specifications Based on Application

State

Source

Specification Category Based on Application

Separation

Stabilization

Drainage

Reinforce-

ment

Paving

Fabric

Alabama 2008 Standard Specifications for

Highway Construction

X X X

Alaska 2004 Standard Specifications for

Highway Construction

X X X X

Arkansas 2003 Standard Specifications for

Highway Construction

X X X X

Arizona 2002 Standard Specifications X X X

California 2006 Standard Specifications X X

Colorado 2005 Standard Specifications

Revision/Special Provision

X X X

Connecticut 2005 Standard Specifications for Roads,

Bridges and Incidental Construction

X

Delaware 2001 Standard Specifications X X X

Florida 2000 Standard Specifications for Road

and Bridge Construction

X X

Georgia 2001 Standard Specifications X X X

Hawaii 2005 Standard Specifications X X X X

Idaho 2004 Standard Specifications for

Highway Construction

X X X

Illinois 2002 Standard Specifications for Road

and Bridge Construction

X

Indiana 2008 Standard Specifications Book X

Iowa 2007 Standard Specifications with

GS.01007 Revisions

X X X

Kansas 2007 Specifications for State Road and

Bridge Construction

X

Kentucky 2004 Standard Specifications for Road

and Bridge Construction

X X

Louisiana 2000 Standard Specifications for Roads

and Bridges

X X X X

Maine 2002 Standard Specifications Revision

for December

X X X X

Maryland 2001 Standard Specifications for

Construction and Materials/Supplement

X X

27

Massachusetts 2007 Standard Special Provisions &

1997 Standard Specs

X X X

Michigan 2003 Standard Specifications for

Construction

X X X

Mississippi 2004 Standard Specs for State Aid Road

and Bridge Construction

X X X X X

Minnesota 2005 Standard Specifications for

Construction

X X X

Missouri 2004 Specifications Book for Highway

Construction

X X X X

Montana 2006 Standard Specs for Road and

Bridge Construction/Supplement

X X X

Nevada 2007 Standard Specifications for Road

and Bridge Construction

X

New

Hampshire

2006 Standard Specifications for Road

and Bridge Construction

X X X

New Jersey 2007 Standard Specifications for Road

and Bridge Construction

X X

New Mexico 2007 Standard Specs for Highway and

Bridge Construction/Supplement

X X X

New York 2006 Standard Specifications X

North Carolina 2002 Standard Specifications X X

North Dakota 2002 Standard Specifications for Road

and Bridge Construction

X X X

Ohio 2008 Construction and Material

Specifications

X X X

Oklahoma 1999 Standard Specifications for

Highway Construction

X X X

Oregon 2008 Standard Specifications/2002

Supplement to Standard Specs

X X X X

Pennsylvania 2000 Standard Specifications for

Highway and Bridge Construction

X X X X

Rhode Island 2007 Standard Specifications for Road

and Bridge Construction

X

South Carolina 2007 Standard Specifications for

Highway Construction/Supplement

X

South Dakota 2004 Standard Specifications for Roads

& Bridges

X X

Tennessee 2004 Standard Specifications for Road

and Bridge Construction

X X

Texas 2004 Standard Specification for

Construction and Maintenance of

Highway

X

Utah 2008 Standard Specifications / 2005

Supplement

X X X

Vermont 2006 Standard Specifications for

Construction Book

X X X

Virginia 2007 Road and Bridge Specifications X X X

Washington 2008 Standard Specifications for Road,

Bridge, and Municipal Construction

X X

Wisconsin 2008 Standard Specifications for

Highway and Structure Construction

X X X

West Virginia 2000 Standard Specifications for Roads

and Bridges

X X X

Wyoming 2003 Standard Specifications for Road

and Bridge Construction

X X X

Total 31 30 42 10 17

Source: Zornberg and Thompson, 2012.

28

The specifications for different geotextile applications indicated that the use for

separation is well established. Further, 31 states had specifications for separation, 30 for

stabilization, and 19 for both.

The accompanying survey indicated that more than one-half of the respondents used

AASHTO M 288 in design. Geotextile selection requirements from various states were

summarized. The survey indicated that the majority of DOTs require AASHTO M 288, Class 2,

geotextile for separation.

At least two states currently use geotextile separators under all pavement types. The

Oklahoma DOT installs separator geotextile under most roadways on the conviction that once

fines have contaminated the aggregate, the drainage function of the aggregate is permanently

compromised (J. Randell, personal communication). In addition, PennDOT is in the process of

implementing separator geotextiles under all pavements (B. Miller, personal communication).

State of the Practice in Virginia

VDOT operates the third largest road network in the United States and is responsible for

maintaining approximately 128,000 lane-miles of interstate, primary, and secondary roads

(VDOT, 2018c). VDOT’s goal, as established by the Commonwealth Transportation Board, is

to have a minimum of 82% of interstate and primary pavements and 65% of secondary

pavements rated as excellent, good, or fair. In 2018, almost 90% of the interstate system was

rated as sufficient, exceeding the statewide performance target of 82%. Approximately 84.7% of

the primary system was rated as sufficient. Statewide, only 60.3% of the secondary road system

had pavement conditions rated as sufficient, effectively not meeting the performance target of

65%. Secondary and subdivision roads comprise approximately 79% of VDOT’s entire road

network. Traditionally, these roads receive comparatively little construction supervision and

service inspection.

Consistent with the practices adopted by other state DOTs, VDOT does not recognize any

structural contribution of geotextiles in pavement design. Current design methodologies include

the MEPDG method for roads carrying traffic volumes in excess of 10,000 vehicles per day

(high-volume roads) and the Vaswani method for secondary and subdivision roads (VDOT,

2018a). Both design approaches suffer from the general lack of recognition of the potential

adverse effect of base aggregate contamination.

The current version of the Vaswani method, as incorporated in VDOT’s Pavement

Design Guide (VDOT, 2018a), contains only two references to geosynthetics. It identifies

geosynthetics as a possible means of mechanical subgrade stabilization to produce a stable

platform for construction equipment. It also states that geosynthetics should be considered for

subgrade stabilization in isolated and limited cases. The guide stipulates that the pavement

section design should be based on the original subgrade soil CBR value, even if some

undercutting and backfilling with dense-graded material is performed to provide a stable

construction platform. The only exception to this rule is if the entire roadway subgrade is

undercut and backfilled to a minimum depth of 2 ft.

29

Some of the documented examples of significant base layer contamination on the VDOT

network include sections of I-81 in Augusta County, as shown in Figure 5. The left photograph

shows subgrade fines on the top surface of the 12-in-thick aggregate base layer. Subgrade soil is

highly plastic clay (CH). The pavement section consisted of approximately 12 in of dense-

graded aggregate and 12 in of asphalt mixture. Poor subsurface drainage led to subgrade

saturation that contributed to this pavement failure.

The right photograph in Figure 5 shows discoloration of the aggregate layer, discovered

during a sinkhole emergency repair spanning the right lane and the shoulder of I-81 in Augusta

County. The degree of contamination was highest near the wheel paths. These examples

indicate that base layer contamination can occur even in a very thick pavement section.

Figure 5. Examples of Base Layer Contamination on I-81 in Augusta County

Survey of VDOT Resident Engineers

A survey was sent to 29 VDOT resident engineers/administrators currently serving in

VDOT’s nine construction districts. Fifteen responses were returned, representing all districts

except Lynchburg and Fredericksburg. The response details are provided in Appendix A, and

the results are summarized as follows.

Approximately 71% of respondents stated that in their experience, subdivision streets that

are accepted into VDOT’s secondary road network fail to reach their 20-year design life. Nearly

43% attributed this failure either to incorrect pavement design factors or to construction quality;

another 7% attributed it partly to those factors but added that the most common cause of failure

was the inaccurate forecast of traffic loading. Two-thirds of respondents stated that subdivision

streets that are accepted by VDOT into the secondary system do not have the same level of

quality as VDOT secondary road projects. Primary causes cited for this perceived deficiency

prominently included inadequate field inspection and oversight. Several respondents specifically

cited unsuitable subgrade being overlooked as the leading factor. Some respondents expressed

30

the opinion that current pavement design requirements for land development / subdivision roads

are inadequate.

The highest cost maintenance activities performed on subdivision streets included asphalt

milling and resurfacing; asphalt overlays; deep culvert and drainage infrastructure replacement;

sidewalk reconstruction; repair of pavement settlements; and mechanized pothole patching and

filling. Typical percentages of maintenance budgets designated for asphalt paving were reported

to range from 1% to 40%. Subdivision streets that had exceeded their design life were most

commonly milled/scarified and overlaid, occasionally with base repair as needed; full

reconstruction was described as rare when mentioned at all.

Of respondents, 60% stated that the majority of subdivision streets in their residency did

not include any form of subsurface drainage system, and 93% stated that contamination of the

aggregate layer by subgrade soil was “sometimes” a contributing cause of pavement degradation

(0 respondents checked “never”). Approximately 73% of respondents stated that they were not

aware of any use of geotextiles in their residency.

Figure 6 summarizes the responses to two questions related to subdivision streets.

Figure 6. VDOT Resident Engineer/Administrator Responses to Survey Questions on Subdivision Roads

VDOT’s Geosynthetic Specifications

Currently, VDOT does not provide any design guidance for using geotextile separators in

road construction. The only specification relevant to pavement structures applies to the

stabilization geotextile, per Section 245 of the Road and Bridge Specifications (VDOT, 2016). It

is typically placed between the undercut subgrade and the replacement granular material. The

stabilization geotextile is also used in a drainage blanket below the pavement section, as a

wrapping fabric around the open-graded aggregate, typically consisting of VDOT No. 2 or No. 3

stone.

71.4%21.4%

7.1%

No Yes Usually yes

In general, is it your experience that

subdivision streets that are accepted

into VDOT's secondary system reach

their 20-year design life?

66.7%

33.3%

No Yes

In your opinion, are subdivision streets

that are accepted into VDOT’s secondary

system of the same level of quality

(i.e. as durable) as VDOT secondary

roadway projects?

31

VDOT subgrade stabilization geotextile is specified to meet the requirements of

AASHTO M 288, Class 3, and the maximum area opening size (AOS) corresponding to the No.

20 sieve (0.850 mm). Presently, no minimum permittivity requirement is specified and any

geotextile type is allowed. As a result, the AOS requirement is satisfied by virtually every

geotextile currently listed in the Industrial Fabrics Association International (IFAI) Specifier’s

Guide (IFAI, 2018). VDOT’s Materials Division compiles a list (Approved List No. 63) of

approved geosynthetic products (VDOT, 2018d). Current VDOT specifications do not stipulate

any installation requirements for geosynthetics. All geosynthetic materials used on the VDOT

road network are required to be NTPEP tested.

VDOT’s Resilient Modulus Database

VDOT’s Materials Division has been collecting laboratory test results on Virginia soils as

part of the resilient modulus testing program set up to facilitate the implementation of the

MEPDG pavement design methodology. In addition to resilient modulus results, various other

tests were performed to characterize and classify these soils. VDOT’s database contains

comprehensive results of 648 laboratory tests performed on 402 subgrade soils, including

gradation, plasticity, water content, and CBR values.

Figure 7 shows the plasticity chart of a subset of Virginia soils containing more than 35%

fines, with CBR values in the range of 3 to 8.

Figure 7. Plasticity Chart for Virginia Soils With 3 < CBR < 8 and More Than 35% Fines. CBR = California

bearing ratio.

32

The bubble size of each test result shown in Figure 7 is proportional to the magnitude of

the CBR value. Most of the high-value CBR soils plot below the liquid limit of approximately

30. It can be seen that Virginia soils containing in excess of 35% fine particles and having CBR

values ranging between 3 and 8 essentially mirror the characteristic pattern of pumping-

susceptible soils identified in Figure 4. Data shown in Figure 7 are representative of

approximately 41% of all soils tested by VDOT. These soils are cohesive according to the

AASHTO soil classification system, as the threshold between the fine-grained and coarse-

grained soils is 35% of particles passing the No. 200 sieve.

Performance of Sites With Geosynthetics in Virginia

VDOT constructed two experimental sections, one in Bedford Country and one in

Albemarle County to assess the structural benefits of using geosynthetics in a paved secondary

road. They have been in service for approximately 24 and 10 years, respectively.

Past Performance

The Bedford County site is well documented. It was evaluated by researchers from

Virginia Tech (Al-Qadi and Appea, 2003, Appea, 1997; Bhutta, 1998; Brandon et al., 1996).

Thirty months into the study, Appea (1997) found a significant difference in the rutting

performance between the control and geotextile sections where the base aggregate layer was 4 in

thick. After analyzing the FWD data, Appea hypothesized that the fouling of the aggregate layer

with the underlying subgrade soil was the main reason for the difference in rutting performance

and that the presence of the geotextile separator layer contributed to the reduction of rutting by

up to approximately 50%. For the subsequent 8 years, the pavement condition was assessed (Al-

Qadi and Appea, 2003) using FWD testing, and the contribution of the geotextile was confirmed

using rutting measurements and apparent subgrade modulus values from the FWD back-

calculation. In addition, three sections with a 4-in base course were excavated to verify

aggregate contamination after 3 years of service; the geotextile section showed the least amount

of contamination (Bhutta, 1998). This section had less rutting than the unreinforced section in

2001 when the base aggregate thickness was 4 in; much of the rutting occurred in the first 3

years. The sections with thicker base courses did not show any significant difference in

performance among the control, geotextile, and geogrid sections.

The Albemarle County site was constructed on a section of Virginia Route 743 near the

Charlottesville-Albemarle Airport. Geotextile was placed at the interface between the aggregate

base and the subgrade in one lane of the section to prevent failure attributable to intermixing, and

the adjacent lane was left unmodified as a control. Statistical analysis of data gathered by an

FWD during construction and after 8 months of traffic showed the lane with the geotextile

section to have a slightly higher structural capacity. As severe intermixing and base failure in

the control section were not expected during the 8-month period because of the low volume of

traffic and a dry season, the benefit may have been attributable to a reinforcing effect because of

construction traffic.

33

Recent Performance

The performance of both sections was evaluated recently using visual observations, an

automated digital video condition survey, FWD testing, and test pit excavation to study the

potential base aggregate contamination.

Bedford County Site

The pavement condition at this site was visually assessed in May 2017. The major

distresses for most of the sections were extensive, full-width transverse cracks, with no visible

differences among any of the nine sections. Figure 8 shows some of the surface distresses.

VDOT’s Maintenance Division collected pavement condition data in 2015 through video

logging and an automated digital condition survey for 100% of VDOT’s secondary roadway

system. The pavement condition rating was calculated based on these video images. This is

based on a deduct value system in which points are deducted from 100 based on the surface

distresses and two ratings are determined: the load related distress rating, composed of pavement

distresses considered to be primarily load related, and the non-load related distress rating,

composed of distresses considered to be primarily non-load related (i.e., climate, materials, or

construction deficiency). The load related distress rating represents the structural adequacy, and

the non-load related distress rating represents the functional condition. Both indices range from

0 for a very poor pavement to 100 for a pavement in perfect condition. VDOT considers values

of 90 and above as excellent and 49 and below as very poor.

Figure 8. Pavement Surface Conditions at the Bedford County Test Site in 2017

34

The pavement condition ratings for each of the nine sections were extracted from the

database and are presented in Figure 9; only the northbound ratings were available. Although

both geotextile and geogrid sections were structurally in better condition (above 90 except for

Section 5) than the control sections for all three base thicknesses, ratings for the controls were

around 80 or better, which is considered adequate. On the other hand, the functional condition

was poor to very poor for all of them, approximately 50 or below, except for Section 9, which

was located on the higher elevation. In contrast with earlier observations (Al-Qadi and Appea,

2003), there was very little rutting (<5 mm) observed in any of the sections during the 2015

survey. It is important to note that these sections were overlaid with HMA at least once in the

last 24 years, estimated to be around 2005 (no detailed maintenance records were kept). A drill

core from each section was collected for thickness measurements. Pavement thicknesses

throughout each section were verified using ground-penetrating radar.

Figure 10 shows the HMA thicknesses for all sections along with the original (Bhutta,

1998) constructed thicknesses in 1994; approximately 2 in of additional overlay on the

northbound lane, tapering to 1 in of overlay on the southbound lane, were placed since the

original construction. As opposed to the significant rutting observed in the first few years

(Appea, 1997), no appreciable rutting was observed with the 2015 rating; the additional overlay

thicknesses and the less severe rain events since 1997 may have contributed to this outcome.

Figure 9. Pavement Condition Rating in 2015 for Bedford County Sections

35

Figure 10. Asphalt Layer Thickness for Nine Bedford County Sections. NB = northbound; SB = southbound.

The structural capacity of these sections was evaluated using the FWD, which is a

standard device for testing the response (deflection) of pavement layers under dynamic load. A

series of velocity transducers (geophones) record the peak surface deflections of the pavement

structure in response to a pulse load supplied by a falling mass and transferred to the pavement

by a steel load plate with a radius of 5.91 in. Four standard load levels (6, 9, 12, and 16 kips)

were used to apply the dynamic load, and deflections were measured at 0, 8, 12, 18, 24, 36, 48,

60, and 72 in from the center of the load plate. FWD deflection data were collected at five points

per section per lane: three at the right wheel path and two at the center. Since there were many

transverse cracks on such short sections, a full basin deflection analysis was not performed. The

simplest comparison would use surface modulus, which is calculated as load over deflection and

is independent of layer thicknesses. Instead of surface modulus, the deflection right below the

load (D0 deflection) was used since the variation in load was not high enough to influence the

modulus value. The average deflections corresponding to 9,000 lb were calculated and

compared statistically (using the t-test); the results are presented in Table 3. Since the overlay

thicknesses in the northbound and southbound lanes were different, they were compared

separately.

In general, the geotextile sections showed lower average deflections than the control

sections (no geosynthetics) under a 9,000-lb load, but there was no statistical difference between

them. Although a lower deflection indicates a stronger structure, the numerical difference is only

about 1 mil or less in most cases except for two southbound thicker-base sections. The

performance of the geogrid sections was inconclusive: Section 9 was very strong and Section 6

was the weakest. During coring, the subgrade soil in Section 9 was found to be dry and stronger

than that of all other sections. It was also at the highest grade elevation.

Historical FWD deflections under the 9,000-lb load immediately after construction in

October 1994 were compiled from the Virginia Tech report by Appea (1997) and plotted along

with the 2017 data in Figure 11 for comparison. The northbound deflection data followed a

trend similar to that observed at initial construction.

36

Table 3. FWD D0-Deflection Data Under a 9,000-Lb Load for Bedford County Site

Base Aggregate

FWD Data: D0 Deflection for 9 Kips Load (mils)

Northbound Southbound

4-in Base Aggregate

Section 1 2 3 1 2 3

Geosynthetics None Geotextile Geogrid None Geotextile Geogrid

Average 13.6 12.8 15.4 31.1 30.0 28.4

Standard deviation 1.99 2.43 1.52 3.30 3.95 3.71

Statistical

significance at 5%

1 vs. 2: No 1 vs. 2: No

2 vs. 3: Yes 2 vs. 3: No

1 vs. 3: No (7%) 1 vs. 3: No

6-in Base Aggregate

Section 4 5 6 4 5 6

Geosynthetics None Geotextile Geogrid None Geotextile Geogrid

Average 28.3 27.7 30.0 31.5 28.2 39.3

Standard deviation 7.00 3.84 3.46 5.65 5.45 5.83

Statistical

significance at 5%

4 vs. 5: No 4 vs. 5: No

5 vs. 6: No (6%) 5 vs. 6: Yes

4 vs. 6: No 4 vs. 6: Yes

8-in Base Aggregate

Section 7 8 9 7 8 9

Geosynthetics None Geotextile Geogrid None Geotextile Geogrid

Average 25.7 24.7 15.6 32.3 27.3 18.3

Standard deviation 3.76 4.70 0.50 7.37 5.04 1.95

Statistical

significance at 5%

7 vs. 8: No 7 vs. 8: No

8 vs. 9: Yes 8 vs. 9: Yes

7 vs. 9: Yes 7 vs. 9: Yes

Figure 11. FWD Deflection (Do) Under a 9,000-lb Load for Bedford County Sections. FWD = falling weight

deflectometer.

37

Some of the sections were excavated to determine if there was base aggregate

contamination with subgrade soil. The bound asphalt layers were saw cut and then carefully

removed in 5 ft × 5 ft sections. Unbound base aggregate samples were collected with hand

digging tools in 2-in increments and then analyzed at the VTRC laboratory for grain size

distribution. In general, all base aggregate gradations, irrespective of depth and location

(Sections 1 through 9), were similar and were mostly within VDOT base aggregate specification

limits, as shown in Figure 12. All aggregate materials had a much higher percentage passing the

No. 200 or No. 100 sieve with respect to the as-constructed gradation, but these same differences

were observed in the 1997 excavation (Bhutta, 1998). Thus, no clear evidence of contamination

was discovered, as the results were similar for all nine sections and occurred during the first 3

years, as documented in previous reports (Appea, 1997; Bhutta, 1998).

Figure 12. Excavated Base Aggregate Gradation: top, Sections 1-3 with 4 in of base aggregate; bottom,

sections with 6 in and 8 in of base aggregate

38

The base samples close to the subgrade layer were observed to have a distinctive clay

stain during wet sieve analysis. With the exception of Section 1, none of the nine sections

showed the presence of any appreciable red clay stain. The Section 1 sample showed a slight red

stain in the wash water. The samples from Sections 1, 2, and 3 were further analyzed using a

hydrometer for grain size distribution of materials passing the No. 200 sieve. The clay-size

particle contents were approximately 4% for Section 1 and 3% for Sections 2 and 3. With the

possible exception of Section 1, these grain size analyses indicated that no significant base

contamination had occurred. Some representative samples were also analyzed using XRD to

detect the presence of underlying clay in the granular base layer.

Fines passing the No. 40 and No. 200 sieves were analyzed for mineral composition

using an Olympus BTX XRD. Diffraction patterns were analyzed using XPowder, Version

2010.01.26 software with the American Mineralogist Crystal Structure Database to identify the

minerals present. Base materials were washed over the No. 40 and No. 200 sieves, and wash

water was dried to prepare the samples. Samples were crushed to ensure that all materials passed

the No. 100 sieve as required by the diffractometer. Samples of subgrade from Section 2 and

Section 5 were analyzed and exhibited similar patterns in which kaolinite and quartz were

identified. Diffraction patterns for the base material showed its basic mineral content to be

calcite, dolomite, and quartz. In the base material analyzed from Sections 2, 3, 4, and 5, calcite,

dolomite, and quartz were the only minerals identified except for the Section 4 sample in which a

trace amount of montmorillonite was identified. In Section 1 base material samples, kaolinite

was identified in addition to calcite, dolomite, and quartz; this observation is consistent with the

red stain observed during the wet sieve analysis mentioned previously and confirms the presence

of some contamination in Section 1.

The exhumed geotextiles were tested for retained tensile strength and permittivity at the

George Mason University geotechnical laboratory (Ullah et al., 2019). Geotextiles exhumed

from below the 4-, 6-, and 8-in-thick base courses retained 68%, 64%, and 60% of their tensile

strengths, respectively. The strains at ultimate tensile strength values showed that there was not a

significant reduction as compared to the strains reported by the manufacturer. The results

indicated that no significant polymer aging took place over the 24 years in service. The

difference in the permittivity values of the original material and cleaned geotextiles showed a

reduction of approximately 60% in the test section with a 4-in base course and 72% in sections

with a 6- and 8-in base course. These permittivity reductions are attributed to the geotextiles

being compressed under the prolonged surcharge pressure, resulting in the decreased pore

openings.

Albemarle County Site

The pavement condition at this site was visually assessed in October 2016 and March

2018. Although the 2016 observation did not reveal any distresses, there were some cracks in

both the northbound and southbound lanes in 2018, as shown in Figure 13.

39

Figure 13. Surface Conditions at Albemarle County Site in 2018

VDOT’s Maintenance Division provided pavement condition data for December 2016.

The data were available only for the northbound lane, which is where the geotextile is located.

In order to compare the performance, data adjacent to the geotextile section in the northbound

lane were used as a surrogate control and are presented in Figure 14. Both the structural (load

related distress rating) and functional (non-load related distress rating) conditions show that the

geotextile and adjacent sections were in very good condition after 10 years of service, with

ratings above 90.

Figure 14. Pavement Condition Rating in 2017 for Albemarle County Site

40

FWD testing, similar to that conducted for the Bedford County sections, was performed

at 30-ft intervals. The resulting data were analyzed with VDOT’s FWD analysis program

MODTAG. To compare the two sections at each test point, the effective structural number and

the effective subgrade resilient modulus of the pavement were computed based on deflections

recorded under a 9,000-lb load. The apparent subgrade resilient modulus for each point was

computed from the deflection readings of the outer sensors so that only the influence of the

subgrade was measured. The effective structural number was computed based on the effective

resilient modulus and center sensor deflection. The deflection (Do-deflection) directly under a

9,000-lb load, effective structural number, and subgrade resilient modulus are shown in Figure

15. There were no statistical differences (at the 5% level of significance) between the

northbound and southbound measurements. One very low deflection recorded toward the middle

of the lane was due to the presence of a drainage structure.

Figure 15. FWD Defection Under 9-kip Load and Back-calculated Effective Structural Number (SNeff) and

Subgrade Modulus (MR) for Albemarle County Site. FWD = falling weight deflectometer.

41

Summary of Findings From Field Investigations

Neither Virginia test site provided a conclusive resolution of the long-term efficacy of

geotextile separators. It should be recognized that neither site was explicitly designed to

evaluate separation. They were used in this study because no adequate long-term records could

be compiled for any other roadway geosynthetic installations on the VDOT network.

Although no visual difference in surface condition was observed between the

geosynthetic and control sections for the Bedford County site after 24 years of service, the actual

distress ratings indicated that all nine sections were in good structural condition (this is expected

given that no contamination was detected), but their functional condition was uniformly very

poor, as shown in Figure 9. All the sections were overlaid with approximately 1 to 2 in of

asphalt around 2005. Both the geotextile and geogrid sections for all base thicknesses exhibited

better structural conditions than their respective control sections, implying that geosynthetics

(geotextiles and geogrids) provided some stabilization benefit or preserved the structural

condition.

The deflections recorded with the FWD equipment at the Bedford County site were lower

for the geotextile section than for the control or geogrid sections for all three base thicknesses,

implying better performance, but the differences were not statistically significant at the 5% level,

as shown in Table 3. The comparative geogrid section performance was inconclusive.

Overall, the Bedford County test site with its relatively high subgrade CBR values did not

provide conclusive proof of the performance benefit attributable to the separation function. The

gradation analyses of base aggregates from the Bedford County sections did not indicate a

significant difference in grain size distribution among any of the nine sections, irrespective of

depth. In addition, there was no evidence of contamination of base aggregate with subgrade clay

except for Section 1, where a trace of red clay was observed in the wash water and in the XRD

analysis. This was likely due to the 4% longitudinal profile grade of the roadway, coupled with

the superelevation, essentially keeping the pavement subbase and subgrade soil relatively well

drained.

There was no measurable difference in performance between the control and geotextile

sections at the Albemarle County (Charlottesville-Albemarle Airport) site after 10 years of

service. One possible reason for this outcome was the very thick pavement structure at this site.

Another is that the edgedrain, constructed along the entire test section, effectively prevented

subgrade saturation, reducing the likelihood of subbase contamination.

Life Cycle Cost Analysis

The LCCA methodology developed in this study combines empirical cost data with a

quantitative pavement deterioration model to measure the cost-effectiveness per mile of

pavements with separator geotextile over a 50-year analysis period. VDOT cost information was

gathered from the SiteManager construction database for the period 2008-2018. Based on

VDOT contract data, an empirical relation between bid quantity and unit bid price was

42

developed for major items so that unit prices would vary with project size, corresponding to

actual VDOT costs. These methods produced price-quantity formulae for separator geotextile;

surface, intermediate, and base asphalt mixtures; aggregate; planing; and guardrail. Quantitative

benefits of separator geotextile were shown through slowed deterioration of aggregate by

subgrade fines relative to pavements without geotextile between the subgrade and subbase, as

discussed in the “Methods” section. The cost data required for the LCCA, the benefits of

separator geotextile, and the results of the LCCA model are discussed in this section.

Costs

The unit cost of greatest uncertainty was that of geotextile used for pavement layer

separation in a roadway. Currently, VDOT does not require that placement of geotextile

separator on subgrade be tracked with a unique item code. Typically, three item codes are used

to track geotextiles purchased by VDOT: 00355 (Geotextile—Subgrade Stabilization), 27500

(Geotextile Fabric), and 00154 (NS Geotextile). At least one additional nonstandard item code is

also used to identify geosynthetic products, including geogrid materials.

Overall, VDOT construction records indicated that item codes 00355 and 27500 were

commonly used for most geosynthetic fabrics and that both codes can denote geotextile

separation function in roadways. The nonstandard item code 00154 is used most often for

geogrid but also for a number of other geosynthetic products: filter cloth, drainage fabric, paving

fabric, nonwoven geotextile fabric, impervious geofabric, membrane, impervious liner, subgrade

stabilization, and high-strength geosynthetic reinforcement. It was also observed that some of

the detailed descriptions of products purchased under (NS) item code 00154 appeared to fit

satisfactorily into either of the standard item codes of 00355 or 27500.

For the purposes of selecting an item code representative of a geotextile used for roadway

separation purposes, it was determined that a geotextile subject to Specification 305 and suitable

for subgrade and shoulders (item code 00355) would be preferred over one subject to

Specification 303 and suitable for earthwork (item code 27500). For this reason, item code

00355 was selected for tracking cost data representative of geotextile used for a separation

function.

Within item code 00355, it was expected that a unit cost estimate for item code 00355

geotextile used in a pavement LCCA would be improved if roadway projects were separated

from other transportation projects. VDOT’s SiteManager database employs a taxonomy for

project type in item code reports, which proved valuable in this regard. The item code report

showed that item code 00355 was used for 21 types of projects over the period 2008-2018,

including 5 types related to asphalt-surfaced roads. These project types were Grade/Drain/Pave,

New Roadway, Pavement Repair, Paving/Asphalt, and Widen Roadway. These 5 project types

were selected to the exclusion of other project types for further unit price analysis of the item

code 00355 geotextile.

Figure 16 shows the total annual quantities of item code 00355 geotextile bid and

installed for these 5 project types combined, as well as annual inflation-adjusted weighted

average unit prices, for the decade 2008-2018.

43

Figure 16. Item Code 00355 (Geotextile—Subgrade Stabilization) in Roadway Projects, 2008-2018 (2018

projects are in progress; quantities are installed, not final). Bid prices are inflation-adjusted weighted

averages of project bid prices. SY = square yard.

It was observed that installed quantities of item code 00355 geotextile typically exceeded

bid quantities over the period shown. Figure 17 shows the total bid and installed quantities of

item code 00355 geotextiles, grouped by each of the five roadway-related project categories for

the decade 2008-2018. It is evident that Grade/Drain/Pave and Widen Roadway project types

dominated bid and installed quantities of item code 00355 geotextiles and for the decade as a

whole in every project type that the quantity of geotextile installed exceeded the quantity that

was bid. Because of their dominance, these categories were developed further in Figures 18 and

19.

Figure 17. Item Code 00355 (Geotextile—Subgrade Stabilization) in Roadway Projects by Project Type, 2008-

2018. Bid prices are inflation-adjusted weighted averages of project bid prices. SY = square yard.

$-

$0.50

$1.00

$1.50

$2.00

$2.50

$3.00

$3.50

-

50

100

150

200

250

300

350

400

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

BID

$ /

SY

QU

AN

TITY

BID

OR

INST

ALL

ED (1

000

SY)

00355 INSTALLED (SY) 00355 BID (SY) BID PRICE ($/SY)

$-

$0.50

$1.00

$1.50

$2.00

$2.50

$3.00

0

200

400

600

800

1000

1200

1400

BID

$ /

SY

QU

AN

TITY

BID

OR

INST

ALL

ED (1

000

SY )

00355 INSTALLED (SY) 00355 BID (SY) BID PRICE ($/SY)

44

Figure 18. Grade/Drain/Pave Project Types, 2008-2018. Bid prices are inflation-adjusted weighted averages

of project bid prices. SY = square yard.

Figure 19. Widen Roadway Project Types, 2008-2018. Bid prices are inflation-adjusted weighted averages of

project bid prices. SY = square yard.

$-

$0.50

$1.00

$1.50

$2.00

$2.50

$3.00

$3.50

$4.00

-

50

100

150

200

250

300

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

BID

($/

SY)

QU

AN

TITY

BID

OR

INST

ALL

ED (1

000

SY)

00355 INSTALLED (SY) 00355 BID (SY) BID PRICE ($/SY)

$-

$0.50

$1.00

$1.50

$2.00

$2.50

$3.00

$3.50

$4.00

-

50

100

150

200

250

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

BID

($/S

Y)

QU

AN

TITY

BID

OR

INST

ALL

ED (1

00

0 S

Y)

00355 INSTALLED (SY) 00355 BID (SY) BID PRICE ($/SY)

45

Figures 18 and 19 show details for the dominant roadway uses of item code 00355

geotextile: Grade/Drain/Pave and Widen Roadway projects. It is noted here that unit prices by

project type were not correlated over the decade, having a Pearson correlation coefficient value

of –0.128.

Within each project type, however, unit prices were at decadal minimums in the years in

which bid quantities most exceeded installed quantities, suggesting a price-steadying effect of

relatively low volumes of change order or overrun purchases of item code 00355. These 2 years

are 2016 in Figure 18 and 2013 in Figure 19. Although Figures 18 and 19 show that bids slightly

exceeded installed quantities in 2012 as well, the margins of 2013 and 2016 suggest a smaller

influence of potentially price-distorting change orders and overruns than in years in which

installed quantities exceeded bid quantities.

The selection of 2013 and 2016 as years from which a fair benchmark unit bid price of

item code 00355 geotextile may be estimated is supported further by the expectation that the

planned use of geotextile as a subgrade separation layer on a low-volume secondary road must

carry close to the lowest installation costs of the possible uses of item code 00355 geotextile.

For these reasons, the quantities bid and unit bid prices in 2013 (for Widen Roadway projects)

and 2016 (for Grade/Drain/Pave projects) are combined to show a relationship between the bid

quantity and unit bid price for item code 00355 geotextile used for separation purposes in

secondary roadway projects. Figure 20 shows the resulting bid quantity-unit price relationship

for item code 00355 geotextile used in roadway projects as developed for this LCCA.

Figure 20. Quantity-to-Price Relationship for Item Code 00355 in Select Roadway Projects (2013 Widen

Roadway and 2016 Grade/Drain/Pave). Bid prices are inflation-adjusted project bid prices. SY = square

yard.

y = -0.471ln(x) + 6.1474

$-

$2

$4

$6

$8

$10

$12

0 20 40 60 80 100 120

BID

PR

ICE

( $ /

SY

)

BID QUANTITY (1000 SY)

GEOTEXTILE (Item Code 00355)

46

Equation 3, expressing the logarithmic trendline of the data shown in Figure 20,

represents the VDOT bid quantity-bid price relationship for item code 00355 geotextile used in

all subsequent cost analyses.

Inflation adjusted unit bid price for item 00355 = −0.471 × ln(Bid quantity) + 6.1474 [Eq. 3]

Bid quantity-to-unit price relationships for other major construction items in secondary

pavements were also estimated from the SiteManager database, as shown in Figure 21. The

methodology used to identify quantity-price curves for the other major items consisted of four

steps: (1) include only projects in which the item was bid in the original contract; (2) include

only contracts categorized in the item code report as one of the five roadway types plus the type

“Surface (Overlay & Treatment)” if applicable; (3) include only contracts executed in the 2008-

2018 period; (4) eliminate high-price “outlier” bids from the data sets if the coefficient of

variation (standard deviation divided by mean) for the item’s unit price data can be reduced to an

arbitrary threshold of 33% when they are excluded, a threshold causing the elimination of only a

few outliers for most construction items. Planing 0-2” was not subject to this rule because of the

number of data points that would have had to be excluded to reach the threshold coefficient of

variation. Planing 0-2” was therefore separated into three functions depending on bid quantity in

square yards as shown in Figure 21(f), (g), and (h). In addition, the secondary road system was

specified in SiteManager reports for bid prices and bid quantities on aggregate, base material,

intermediate layer, planing, and guardrail. Except for guardrail, which spanned seven item

codes, the quantity-price relationships employed one to three item codes for each item.

The resulting bid quantity-to-unit price scatterplots for all items except planing were

coherent enough to use simple logarithmic trendlines as functional quantity-price relationships.

In the case of planing, price extremes appeared to be the norm at low quantities and thus were

not eliminated.

Benefits

The LCCA framework developed in this study is based on practical assumptions and

actual cost data, leading by logical steps to estimate the cost-effectiveness of geotextile separator

per mile of pavement. The key assumption allowing quantification of results is drawn from the

literature, i.e., that the primary benefit of separator geotextile is the prevention of contamination

of the unbound aggregate layer. Thus, in the resulting LCCA calculations, pavements without

geotextile deteriorate because of the loss of effective subbase thickness whereas pavements with

geotextile do not deteriorate from this cause. Through the application of a pavement

deterioration model that enables a pavement layer rating as shown in Appendix B, the selection

and timing of maintenance activities performed on the pavements with and without geotextile

were determined. The panel of LCCA results include pavement designs and maintenance

schedules for (1) two initial AADT levels; (2) the presence or absence of geotextile separator

between pavement subgrade and subbase; (3) the performance or nonperformance of preventive

maintenance, (4) three contamination rates of aggregate base by subgrade fines; and (5) two

discount rates consisting of a rate commonly assumed in VDOT LCCA and a rate dictated by

current OMB guidance.

47

Figure 21. Bid Price-Quantity Relationships for Major Construction Items in LCCA. All bid prices are adjusted for inflation. LCCA = life cycle cost

analysis; LF = linear foot; SY = square yard.

y = -7.202ln(x) + 143.37

$-

$100

$200

$300

$400

$500

$600

0 20 40 60 80

BID

$ /

TO

N

BID QUANTITY (1000 TONS)

ASPHALT SURFACE MIX (SM-9.5A)

y = -11.03ln(x) + 167.54

$0

$50

$100

$150

$200

$250

$300

0 5 10 15 20 25

BID

$ /

TO

N

BID QUANTITY (1000 TONS)

ASPHALT INTERMEDIATE MIX (IM-19.0A & 19.0D)

y = -10.95ln(x) + 172.53

$-

$50

$100

$150

$200

$250

$300

0 10 20 30 40

BID

$ /

TO

N

BID QUANTITY (1000 TONS)

ASPHALT BASE MIX (BM-25.0A)

y = -2.081ln(x) + 41.065

$0

$20

$40

$60

$80

0 10 20 30 40 50 60

BID

$ /

TO

N

BID QUANTITY (1000 TONS)

AGGREGATE BASE: VDOT NO. 21B

y = 4E-05x2 - 0.0617x + 31.264

$-

$20

$40

$60

$80

0 200 400 600 800 1000

BID

$ /

SY

BID QUANTITY (SY)

PLANING 0" TO 2" (< 1000 SY)

y = -2.023ln(x) + 35.138

$-

$20

$40

$60

$80

0 2000 4000 6000 8000

BID

$ /

LF

BID QUANTITY (LF)

GUARDRAIL

y = 11.884x-0.156

$0

$2

$4

$6

$8

$10

$12

$14

5 205 405 605 805 1005

BID

$ /

SY

BID QUANTITY (1000 SY)

PLANING 0" - 2" (≥ 5000 SY)

y = 877.57x-0.668

$-

$2

$4

$6

$8

$10

$12

$14

0 2000 4000 6000

BID

$ /

SY

BID QUANTITY (SY)

PLANING 0" - 2" (1000 - 4999 SY)

y = -0.528ln(x) + 8.8558

$-

$1

$2

$3

$4

$5

$6

$7

$8

0 100 200 300 400 500 600

BID

$ /

SY

BID QUANTITY (1000 SY)

PLANING 2" - 4"

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

48

LCCA Results

LCCA results are provided in the four tables of Figure 22. The LCCA model determines

life cycle costs for a hypothetical mile of roadway with two 12-ft travel lanes. For 1,500 AADT

only, 2-ft shoulders are provided on each side of the roadway. Pavement design alternatives for

each initial AADT are differentiated by the presence or absence of separator geotextile and of

preventive maintenance. As noted previously, pavement deterioration is modeled according to

the deterioration curves in Appendix B: specifically, surface mixture deteriorates more rapidly

over base mixture with a structural coefficient below 0.33. Each pavement design alternative is

subjected to three contamination rates. Each complete scenario is tested at two discount rates.

Cost results expressed in present value are calculated for a short-term period of 20 years

and for a 50-year analysis period, the latter employing two different treatments of the estimated

remaining roadway service life at the end of the analysis period. In the first treatment, monetized

salvage value is determined in accordance with Equation 4:

Monetized salvage value = (Estimated remaining design life (yr) 20 years⁄ ) ×Replacement cost at 50 years [Eq. 4]

The replacement cost at 50 years was uniform across all pavements with the same initial

AADT and is typical for a secondary road with traffic levels determined for year 50. Monetized

salvage value offsets the discounted stream of maintenance costs over the 50-year analysis

period and reduces life cycle costs in different amounts.

In the second treatment, the “normalized” liability approach, each pavement is given

warranted maintenance at year 50, but each pavement receives a 20-year terminal pavement

design based on its condition in year 50. Deficiencies are billed as hypothetical net costs

required to normalize the pavement with respect to other pavement alternatives for an additional

20 years of service. The net costs of this hypothetical terminal pavement design are discounted

and added to the sum of life cycle costs in a manner identical to the costs of earlier maintenance

activities.

Figure 22 shows the life cycle costs of each scenario in columns by contamination rate

and, in the two bottom (italicized) rows of each table, the calculated savings or added expense of

the geotextile separation layer in the pavement, given the presence or absence of preventive

maintenance activities for both pavements, for a 20-year interim period and for the longer

analysis periods. Minimum life cycle costs among the pavement designs within a scenario (in

each column) are shown by shaded cells; the cost deltas of a pavement with a separator

geotextile layer in each scenario, given a preventive maintenance decision applied to both

pavements, are shown in the two bottom rows.

The results indicated that the 20-year cost-effectiveness of geotextile increases with the

initial AADT and becomes most substantial in the scenario of the higher initial AADT, the high

contamination rate, and the lower discount rate. Geotextile is never cost-effective within 20

years of service life if the contamination rate is relatively low (e.g., 0.05 in/yr), the initial AADT

is very low, and preventive maintenance is foregone.

49

Figure 22. Life Cycle Costs per Mile for Very Low Initial AADT and Low Initial AADT at Two Discount

Rates. AADT = annual average daily traffic; LCC = life cycle cost; GTX = separator geotextile; PM =

preventive maintenance. Shaded cells indicate minimum life cycle costs among pavement designs within a

scenario.

The results also consistently indicated that 50-year life cycle costs for pavements with

geotextile were lower than for roads without geotextile at high- and mid-level contamination

rates (0.2 and 0.1 in/yr), regardless of initial AADT or performance of preventive maintenance

(on both or neither pavement), the discount rate, and the treatment of salvage value. On the other

hand, No GTX-No PM roads were consistently more cost-effective at the lowest contamination

rate with one exception that occurred in the low initial AADT, low contamination rate, low

discount rate, and monetized salvage value approach scenario (life cycle costs of $335,129). The

0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr

GTX-PM 403,994$ 403,994$ 403,994$ 420,646$ 420,646$ 420,646$ 467,873$ 467,873$ 467,873$

GTX-NO PM 385,911$ 385,911$ 385,911$ 365,981$ 365,981$ 365,981$ 434,381$ 434,381$ 433,880$

NO GTX- PM 486,929$ 377,277$ 377,277$ 822,238$ 552,080$ 430,991$ 845,935$ 586,061$ 456,489$

NO GTX - NO PM 468,847$ 359,195$ 359,195$ 738,488$ 428,514$ 356,212$ 776,801$ 459,761$ 394,607$

PM (82,935)$ 26,717$ 26,717$ (401,592)$ (131,434)$ (10,345)$ (378,062)$ (118,188)$ 11,384$

NO PM (82,936)$ 26,716$ 26,716$ (372,507)$ (62,533)$ 9,769$ (342,420)$ (25,380)$ 39,273$

(a) Initial AADT = 155, discount rate = 0.04

0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr

GTX-PM 851,221$ 851,221$ 851,221$ 1,186,732$ 1,186,732$ 1,186,732$ 1,306,328$ 1,306,328$ 1,306,328$

GTX-NO PM 823,152$ 823,152$ 823,152$ 1,124,058$ 1,124,058$ 1,128,309$ 1,233,732$ 1,227,041$ 1,231,292$

NO GTX- PM 1,090,331$ 930,624$ 877,394$ 1,574,374$ 1,247,701$ 1,197,869$ 1,626,952$ 1,352,033$ 1,252,976$

NO GTX - NO PM 1,090,331$ 859,532$ 793,304$ 1,574,374$ 1,259,785$ 1,104,134$ 1,626,952$ 1,364,116$ 1,200,732$

PM (239,110)$ (79,403)$ (26,173)$ (387,642)$ (60,969)$ (11,137)$ (320,624)$ (45,705)$ 53,352$

NO PM (267,179)$ (36,380)$ 29,848$ (450,316)$ (135,727)$ 24,175$ (393,220)$ (137,075)$ 30,560$

(b) Initial AADT = 1500, discount rate = 0.04

0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr

GTX-PM 450,490$ 450,490$ 450,490$ 482,021$ 482,021$ 482,021$ 641,447$ 641,447$ 641,447$

GTX-NO PM 421,072$ 421,072$ 421,072$ 335,129$ 335,129$ 335,129$ 566,026$ 566,026$ 564,336$

NO GTX- PM 581,725$ 641,447$ 423,773$ 1,296,485$ 818,894$ 549,745$ 1,376,476$ 933,604$ 635,820$

NO GTX - NO PM 552,308$ 482,021$ 394,356$ 1,187,205$ 558,037$ 365,621$ 1,316,536$ 663,516$ 495,231$

PM (131,235)$ (190,957)$ 26,717$ (814,464)$ (336,873)$ (67,724)$ (735,029)$ (292,157)$ 5,627$

NO PM (131,236)$ (60,949)$ 26,716$ (852,076)$ (222,908)$ (30,492)$ (750,510)$ (97,490)$ 69,105$

(c) Initial AADT = 155, discount rate = 0.015

0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr 0.2-in/yr 0.1-in/yr 0.05-in/yr

GTX-PM 910,072$ 910,072$ 910,072$ 1,592,934$ 1,592,934$ 1,592,934$ 1,996,657$ 1,996,657$ 1,996,657$

GTX-NO PM 864,408$ 864,408$ 864,408$ 1,445,408$ 1,445,408$ 1,453,218$ 1,815,638$ 1,793,051$ 1,800,861$

NO GTX- PM 1,289,428$ 1,048,956$ 971,362$ 2,388,717$ 1,760,009$ 1,630,425$ 2,566,207$ 2,112,203$ 1,816,451$

NO GTX - NO PM 1,289,428$ 933,300$ 834,561$ 2,388,717$ 1,756,705$ 1,448,295$ 2,566,207$ 2,108,898$ 1,774,380$

PM (379,356)$ (138,884)$ (61,290)$ (795,783)$ (167,075)$ (37,491)$ (569,550)$ (115,546)$ 180,206$

NO PM (425,020)$ (68,892)$ 29,847$ (943,309)$ (311,297)$ 4,923$ (750,569)$ (315,847)$ 26,481$

(d) Initial AADT = 1500, discount rate = 0.015

Discounted Cost @ 20-yrLCC @ 50 years, Monetized Salvage

Value

LCC @ 70 years, Terminal Pavement

Design

Discounted Cost @ 20-yrLCC @ 50 years, Monetized Salvage

Value

LCC @ 70 years, Terminal Pavement

Design

Discounted Cost @ 20-yrLCC @ 50 years, Monetized Salvage

Value

LCC @ 70 years, Terminal Pavement

Design

Discounted Cost @ 20-yrLCC @ 50 years, Monetized Salvage

Value

LCC @ 70 years, Terminal Pavement

Design

50

explanation for this exception is the fortuitous combination of the extremely low cumulative cost

of the GTX-No PM road over the analysis period, a high remaining service life at 50 years, and

the monetized salvage value approach to salvage value incorporation. This exception disappears

under the terminal pavement design approach to the treatment of salvage value.

DISCUSSION

There is an abundance of successful case histories demonstrating that geosynthetics can

be used to resolve constructability problems on very soft soils. They can help establish a

working platform quickly and efficiently while minimizing the loss of base course material

during construction. Significant cost savings can be realized in the initial construction phase by

the proper selection and application of geosynthetic materials, with readily apparent and almost

immediate results.

Although geotextiles were introduced approximately 50 years ago, both researchers and

practitioners have remarked on the absence of published documentation quantifying their long-

term benefits (Christopher, 2014; Holtz, 2016). Koerner (1997) stated that the only justification

for the use of geotextile on firm subgrades is the longer service life of pavements. He proposed a

20-year study designed to analyze the performance of comparable test sites with and without

separator geotextiles to make a proper assessment of their respective service life. The need for

such a study still exists today. The empirical proof of reduced long-term maintenance costs of

pavements with geotextile separators remains to be documented. Unfortunately, there is

insufficient information as to the current status or even a standardized method of assessing the

situation (R. Koerner, personal communication). Reliable data needed for evaluating existing

geotextile pavement sections are virtually nonexistent at VDOT, as the current PMS is not

configured to capture geosynthetic materials used in construction or rehabilitation.

The difference between the length of the initial service life used in many state LCCAs

and the longer actual service lives reported by states in the LTPP database as documented for

asphalt pavements in the NCAT report by Robbins and Tran (2018) indicates that LCCAs in

current use often do not represent actual conditions for highway agencies. Until uniform and

universal physical criteria are adopted in practice, as Robbins and Tran (2018) recommended, an

LCCA model with reasonable assumptions, sufficient detail to capture the function of separator

geotextile, and accurate cost data remains the most effective decision support tool available.

The LCCA model developed in this study provides an analytical assessment tool and a

pathway to future research efforts. Figure 23 condenses the LCCA results from Figure 22 for the

50-year analysis period using both methods of incorporating remaining service life at the end of

the analysis period. Percentage figures represent the cost difference between the GTX pavement

and the No GTX pavement relative to the cost of the No GTX pavement. Unshaded cells show

the cost advantage or disadvantage of the GTX pavement (columns) relative to No GTX

pavements (rows) when both or neither received preventive maintenance over the 50-year

analysis period. Shaded cells show cross-comparisons of pavements; i.e., only one pavement

contains geotextile or received preventive maintenance.

51

In the highest and middle contamination rate scenarios [Figure 23, columns (a) through

(d)], whether or not any pavement received preventive maintenance, the GTX pavements were

consistently less costly in present value than No GTX pavements over the 50-year analysis

period. As expected, the effect of the 1.5% discount rate versus the 4% rate was to increase the

relative cost advantage of GTX pavements. Further, direct comparisons of GTX and No GTX

pavements (both receiving preventive maintenance or neither receiving preventive maintenance,

indicated by unshaded squares) in the mid- to high-contamination rate scenarios showed

negligible changes that were attributable to preventive maintenance in the consistent GTX

pavement cost advantage. It should be kept in mind, however, that prospective cost savings will

come at the price of deterioration in surface condition state and ride quality if preventive

maintenance activities are not performed.

Figure 23. Summary of Life Cycle Cost Analysis Results. AADT = annual average daily traffic; GTX =

geotextile; PM = preventive maintenance. Shading indicates cross-comparisons of pavements; i.e., only one

pavement contains geotextile or received preventive maintenance.

(a) (b) (c) (d) (e) (f)

GTX-No PM GTX-PM GTX-No PM GTX-PM GTX-No PM GTX-PM

No GTX-PM 45% 51% 66% 76% 85% 98%

No GTX-No PM 50% 57% 85% 98% 103% 118%

No GTX-PM 71% 75% 90% 95% 94% 99%

No GTX-No PM 71% 75% 89% 94% 102% 107%

No GTX-PM 26% 37% 41% 59% 61% 88%

No GTX-No PM 28% 41% 60% 86% 92% 132%

No GTX-PM 61% 67% 82% 91% 89% 98%

No GTX-No PM 61% 67% 82% 91% 100% 110%

(a) (b) (c) (d) (e) (f)

GTX-No PM GTX-PM GTX-No PM GTX-PM GTX-No PM GTX-PM

No GTX-PM 51% 55% 74% 80% 95% 102%

No GTX-No PM 56% 60% 94% 102% 110% 119%

No GTX-PM 76% 80% 91% 97% 98% 104%

No GTX-No PM 76% 80% 90% 96% 103% 109%

No GTX-PM 41% 47% 61% 69% 89% 101%

No GTX-No PM 43% 49% 85% 97% 114% 130%

No GTX-PM 71% 78% 85% 95% 99% 110%

No GTX-No PM 71% 78% 85% 95% 101% 113%

Contamination Rate

0.2-in/yr 0.1-in/yr 0.05-in/yr

Discount

Rate

Initial

AADT

Pavement

Design

4.0%

155

1500

1.5%

155

1500

(a) Monetized Salvage Value Method

(b) Terminal Pavement Design Method

Contamination Rate

4.0%

155

1500

1.5%

155

1500

0.2-in/yr 0.1-in/yr 0.05-in/yr

Discount

Rate

Initial

AADT

Pavement

Design

52

For the lowest contamination rate [columns (e) and (f)], the monetized salvage value

approach shows slightly more variety in the relative cost-effectiveness of GTX-PM pavements

than does the terminal pavement design approach. In the latter, GTX-PM pavements [column

(e)] are never cost-effective at the lowest contamination rate. The two approaches have similar

mixed results regarding the cost-effectiveness of GTX-No PM pavements [column (f)]. For the

lowest contamination rate, the tables agree that the largest cost advantage of a GTX pavement, at

-39% in Figure 23(a) and -11% in Figure 23(b), is in the cross-comparison of GTX-No PM vs.

No GTX-PM pavements under very low initial AADT and the lower discount rate. The tables

conflict regarding the scenario with the highest cost disadvantage of a GTX pavement, although

it is clear that the cost discrepancies are larger for pavement designed for very low AADT than

for pavement designed for higher AADT.

Under monetized salvage value [Figure 23(a)] and the higher discount rate, direct

comparisons of GTX and No GTX pavements under the lowest contamination rate showed

similar costs (±3 percentage points around 0) for GTX and No GTX pavements when they both

received or did not receive preventive maintenance, regardless of initial AADT. Yet the GTX

pavement showed an 8% to 12% cost advantage whether both pavements did or did not receive

preventive maintenance in the very low AADT–low contamination rate scenario under the lower

discount rate. This exception was likely due to the low contamination rate and initial traffic

loading, allowing for very low maintenance costs and a high remaining design life by year 50, as

well as the monetized salvage value approach at the end of the analysis period and the low

discount rate. By contrast, use of the terminal pavement design factor indicated that GTX

pavement has no significant cost-effectiveness to offer in this lowest contamination scenario or

any other direct comparison (i.e., both pavements receive or forego preventive maintenance).

GTX pavement showed a cost advantage only if the other pavement received preventive

maintenance in the scenario of the lower discount rate and lower initial AADT.

Overall, the LCCA results presented in this study consistently support the use of

separator geotextile when the estimated contamination occurs at a rate of at least 0.1 in per year,

even when the initial traffic volume is very low and regardless of whether preventive

maintenance is performed. The results indicated that the benefit of separator geotextiles

decreases with a decreasing contamination rate. For the highest contamination rate but

regardless of initial AADT and discount rate, GTX pavement had a consistent and sizable cost

advantage over No GTX pavement when both pavements received preventive maintenance, and

the cost advantage tended to increase if neither pavement received preventive maintenance. The

LCCA results also suggested that the contamination range tested in the LCCA model spans an

appropriate range for defining the contamination parameter that delineates cost advantages from

no cost advantages attributable to geotextile.

The price-quantity relationships determined from VDOT construction records for item

codes 00355 (geotextile) and 10128 (21B aggregate) allow another perspective regarding the

potential cost-effectiveness of using separator geotextile in road construction. Assuming a

hypothetical 100-home subdivision project consisting of 2 miles of 25-ft-wide road,

approximately 29,300 square yards of separator geotextile would be required at a unit cost of

about $1.30 per square yard (from Figure 21). If 100 houses were constructed, the addition of

separator geotextile would result in a delta cost of approximately $380 per house. The study by

53

Al-Qadi and Appea (2003) implied that the life of the subdivision pavement could be nearly

doubled for this delta cost.

When a smaller area is considered, such as a 1,000-ft section of road terminating in a cul-

de-sac with 10 homes, the delta cost for separator geotextile would be approximately $670 per

house. Therefore, using current VDOT data, separator geotextile appears to be most cost-

effective on large projects. The potential benefits of using a geotextile separator can be

identified with reference to Figure 24.

The primary purpose of a geotextile separator layer is to prevent the intermixing of base

aggregate and subgrade fines. It also provides lateral confinement, potentially leading to the

improved durability of the pavement section. Although the intermixing process is more likely to

occur on low-volume secondary roads than on heavily traveled arteries, the outcome is largely

dependent on the actual strain that develops at the subgrade level in response to the applied

traffic loads. In addition, the rate of fines migration can change over time because of a potential

decrease in subbase permeability. Ultimately, the main reasons for using separation fabrics

include preserving the integrity of the unbound aggregate base layer and maintaining the long-

term pavement performance. Although the surficial HMA layer must be replaced periodically as

it wears out under traffic, the base material can be reused in perpetuity if it remains structurally

intact.

It may be argued that the underlying mechanism of base course contamination and

subsequent weakening is in fact significantly more complex than originally advanced and the

actual field outcome is controlled by other inputs. Pumping-susceptible soils are only a part of

the overall phenomenon resulting in pavement subbase contamination. As Woods and Shelburne

observed in their 1943 study, pumping becomes a serious problem only as a result of the

synergistic influence of several factors. Subgrade saturation, porous base course, and heavy

cyclic loading, in addition to pumping-susceptible subgrade soils, must be simultaneously

present. It should be recognized that these same factors contribute to a large percentage of

various types of pavement failures. Absent any one of these components, the conditions for

subbase contamination are greatly diminished, reducing the need for separator membrane. Thus,

subbase contamination may be viewed as a transient phenomenon rather than a continuous

process.

Figure 24. Separator Geotextile in a Pavement System

54

The literature review and field observations indicated that subgrade separator geotextile

is not necessary for all roadways. Generally, the need for separator geotextile may be assessed

based on the following four criteria:

1. Fines content above 35%. This is the breakpoint between granular and silt-clay

materials according to the AASHTO soil classification system. There is no evidence

justifying the need for separation in subgrade soils with the fines content below 35%.

2. Subsurface drainage. This is a key factor to long-term pavement performance.

Simply stated, if subgrade soils do not become saturated, they do not pump. In the

case of VDOT, there are no requirements for any subsurface drainage for low-volume

roads (carrying less than 1,000 vehicles per day). The vast majority of subdivision

roadways within VDOT’s maintenance network fall into this category. Both test sites

(Bedford County and Albemarle County) appear to have resisted contamination

because of adequate subsurface drainage.

3. Subgrade soil strength. Subgrades with soaked CBR values between 3 and 8 should

be considered for the use of separator geotextiles. Subgrade soils identified for the

use of nonwoven geotextile separators exhibit the material properties shown in Figure

7. Stabilization applications should be evaluated below CBR 3.

4. Pavement section design. Long-term pumping of fines is not critical in pavement

structures with a relatively low traffic-induced strain at the subgrade level. The

Albemarle County test site appears to be an example of this.

The concept of adequate subsurface drainage is challenging to define for the multitude of

potential pavement design scenarios. However, the following list (in descending order of

effectiveness) identifies VDOT’s current subsurface drainage design options (VDOT, 2018a,

2018b):

1. Open-graded drainage layers. When connected to subsurface drains, these asphalt- or

cement-stabilized pavement layers have the greatest permeability. Although these

layers also require low-permeability materials underneath and tend to be more

expensive to construct (usually reserved for high-volume roadways), they provide

subsurface drainage well above that required for the typical rainfall in Virginia.

2. Subsurface drains. When connected to dense-graded aggregate layers, a variety of

VDOT standard drains improve the permeability of the pavement structure. Drains

that incorporate perforated pipe have a permeability an order of magnitude higher

than those that use prefabricated slot drains. Analyses using FHWA’s DRIP software

indicate that subsurface drains improve the pavement system’s permeability, but their

effectiveness is dependent on the aggregate layer gradation. Combination

underdrains should be used in all vertical (profile) sag sections to prevent ponding of

water within the aggregate layers; these are also typically used at cut and fill

transitions.

55

3. Daylighted aggregate layers. Simply constructing aggregate layers through the

shoulder and embankment to daylight provides the lowest permeability at the lowest

cost in most scenarios. However, as VDOT uses dense-graded base aggregates and

drainage paths may become exceedingly long, the permeability of the daylighted

layer may not be sufficient in all cases to prevent subgrade saturation. This option is

also the most susceptible to fines contamination by the subgrade soils.

The key to the consistent use and wider adoption of geosynthetics in road construction is

to implement design based on specifications, modeled on the AASHTO M 288 framework. It

should be recognized that the material strength is the primary consideration for durability. There

is an established consensus to require Class 1 material for stabilization and Class 2 material for

separation applications, as outlined previously. The other parameters critical for adequate

performance are AOS and permittivity. These parameters involve a trade-off between the

drainage and filtration properties. The reduction in AOS will ensure the filtration of fine

particles but may significantly impede flow through geotextile, potentially resulting in a

saturated subgrade. The permittivity of a geotextile should be maximized to provide adequate

drainage and redundancy over time.

The selection of appropriate permittivity and AOS parameter values may be optimized to

provide adequate functionality for typical applications. In the case of Virginia, VDOT Materials

Division records indicate that statewide, most of the cohesive subgrade soils contain in excess of

50% fines (approximately 60% of all soil samples tested contain 50% or more fines). Thus, the

optimal permittivity of 0.1 sec-1 may be adopted for design specifications. This requirement

matches the AASHTO M 288 minimum permittivity requirement for erosion control fabrics used

with soils containing 50% or more fines.

The optimal AOS for geotextile is determined based on the soil retention criteria

(Koerner, 2013). Sufficient conditions to prevent soil loss through the geotextile exist when the

soil contains more than 30% fines, has a coefficient of permeability less than 1 x 10-7 cm/s, has a

plasticity index greater than 15, and has an undrained shear strength in excess of 200 psf. A

geotextile with an AOS value less than that of a No. 70 sieve (0.212 mm) can be used with soils

meeting these criteria, essentially functioning more as a separator than a filter. This AOS

requirement can be considered applicable to the vast majority of cohesive subgrade soils in

Virginia.

Proposed application guidelines for the use of roadway geosynthetics are presented in

Table 4. These guidelines integrate current VDOT practices with respect to the treatment of

unsuitable soils and build on the findings of this study as related to subgrade separation.

Subgrade stabilization guidelines indicate that this practice should be viewed as an alternative to

chemical stabilization or excavation and replacement.

56

Table 4. Proposed Guidelines for Use of Roadway Geosyntheticsa

Primary

Application

(Pay Item)

AASHTO M 288

Class and Material

Type

Minimum

Permittivity

(sec-1)

Maximum

AOS

(mm)

Guidelines for Useb

Subgrade

Stabilization

Class 1

No silt-film woven

fabrics allowed.

Biaxial geogrid can be

used with nonwoven

or woven separator

geotextile

0.1 0.212

(No. 70)

Any of the following:

CH, MH, OH, OL, PT (VDOT,

2019)

organic content > 5% (VDOT,

2019)

swelling > 5%

natural WC > 30% above OWC

(VDOT, 2005)

subgrade CBR < 3 (AASHTO,

2017)

CBR or MR < design value

as judged by the Engineer (VDOT,

2013, 2019)

Subgrade

Separation

Class 2

Nonwoven geotextile

only

0.1 0.212

(No. 70)

All of the following:

more than 35% subgrade fines

(Kroening et al., 2017a)

subgrade CBR between 3 and 8

(AASHTO, 2017)

no adequate pavement subdrainage

as judged by the Engineer

AOS = apparent opening size; CH = fat clay; MH = elastic silt; OH = organic clay or silt; OL = organic clay or silt;

PT = peat; WC = water content; OWC = optimum water content; CBR = California bearing ratio; MR = resilient

modulus. a Although the use of stabilization geosynthetic(s) can reduce the depth of undercut and quantity of replacement

material, and in some limited cases eliminate them, VDOT pavement design methods should be followed without

adding any structural benefit of the geosynthetic(s). b

Special considerations: subgrade separator geotextile should be used between any open-graded aggregate layer and

soil and for all jointed concrete pavements constructed on pumping-susceptible subgrades.

The proposed Special Provision for Subgrade Separation and Subgrade Stabilization

Geotextiles is presented in Appendix E. It is estimated that more than 70 currently available

nonwoven geotextiles will meet the new separation criteria, based on the IFAI Specifier’s Guide

(IFAI, 2018).

Accurate determination of the “expected contamination rate” of an unknown roadway is a

very challenging problem. A practical solution is to ensure that contamination does not occur,

using a cost-effective analytical approach. Guidance provided in this study, coupled with LCCA

modeling, is focused on minimizing subbase contamination, leading to optimal life cycle road

maintenance expenditures.

CONCLUSIONS

The majority of state DOTs use geosynthetics in distinct separation and stabilization

applications.

57

Geosynthetic materials can provide long-term benefits when applied as an integral part of

the pavement design practice without reducing the aggregate base or bituminous layer

thickness.

Separator geotextiles are most suited to relatively thin pavement sections constructed on

pumping-susceptible soils and roadways with or subject to inadequate subsurface drainage.

The optimal range of soaked CBR values applicable to geotextile separators is between 3 and

8. AASHTO M 288 identifies a CBR of 3 as the threshold value between separation and

stabilization.

Separator geotextiles can help maintain the integrity of the unbound base material under

some conditions, typically representative of subdivision and secondary roads. The base can

remain functional in perpetuity if its structural integrity is not compromised by the intrusion

of subgrade fines.

Subgrade stabilization using geosynthetics is most applicable to soft soils with soaked CBR

values below 3. Site-specific stabilization design is required for subgrade soils with CBR

values below 1.

The results of the survey of VDOT resident engineers indicate the need to consider potential

quality improvements in the construction of subdivision roads, although it is not evident that

all of the identified problems could have been prevented by the use of geosynthetics.

Current record keeping at VDOT does not lend itself to a detailed empirical analysis of

geosynthetics use on Virginia roadways because no geotextile item code specifies a pavement

layer separation function. In addition, VDOT’s PMS does not currently document

geosynthetics used in construction or rehabilitation.

The Virginia test sites did not corroborate the claim of long-term benefits with woven

geotextiles and geogrids. Adequate subdrainage has kept these subgrade soils from

becoming saturated. Pavement section overdesign (Albemarle County site) and maintenance

overlay placement (Bedford County site) may have masked the identification of potential

benefits.

The LCCA model developed in this study permits differentiation and quantification of the

benefits of using separator geotextile under varying subbase contamination rates for a low-

volume roadway under two AADT assumptions and across a 50-year time period. The

results of the LCCA model indicate the following:

The 20-year cost-effectiveness of geotextile increases with the initial AADT assumption

and becomes substantial in the scenario of the higher initial AADT level, a high

contamination rate, and the lower discount rate, although geotextile is not cost-effective

within 20 years if the contamination rate is as low as 0.05 in/yr, initial AADT is at the

lower level, and preventive maintenance is not provided.

58

Pavements with separator geotextile are consistently cost-effective compared to

pavements without separator geotextile at contamination rates above 0.1 in/yr,

regardless of whether preventive maintenance is performed or whether initial AADT is

set at the lower or higher level, for discount rates within the range explored in this study

(4% and lower).

Separator geotextile in pavement has a consistent and sizable life cycle cost advantage

over pavement without geotextile for the highest contamination rate tested in this study

when both pavements receive preventive maintenance, and the cost advantage usually

increases if neither pavement receives preventive maintenance, regardless of initial

AADT.

The life cycle cost advantage of separator geotextile decreases with a decreasing

contamination rate.

Judicious use of separator geotextile can increase the reliability of unbound base aggregate

at a relatively low unit cost on large construction projects at current VDOT bid prices.

RECOMMENDATIONS

1. VDOT’s Materials Division should revise current geosynthetic specifications to include

“Subgrade Stabilization” and “Subgrade Separator Geotextiles” as two separate and

distinct pay items. This will result in providing additional design options and better tracking

of geosynthetics use on VDOT roadways. The addition of the subgrade separator geotextile

option is expected to result in a more economical use of fabrics on VDOT roadways by

differentiating textile properties and applications. It is also expected to help maintain

performance of unbound subbase material.

2. VDOT’s Materials Division should adopt the guidelines presented in Table 4 for the use of

geosynthetics on low-volume roads. This will result in optimized pavement design for low-

volume and subdivision roadways. The use of these geosynthetics consistent with the

guidance provided is expected to reduce VDOT maintenance and pavement replacement

costs.

3. VDOT’s Materials Division should implement changes to the existing geosynthetics special

provision, as presented in Appendix E. The proposed modifications expand on the existing

stabilization application and provide additional options to the road designer.

4. VDOT’s Maintenance Division should update PMS software to allow the inclusion of

geosynthetics in its pavement structure database. This will result in the documentation of the

use of geosynthetics in pavements, allowing future long-term studies of pavement

performance.

59

SUGGESTIONS FOR FURTHER RESEARCH

The literature review highlighted a particular need for additional research on the long-

term benefits of geosynthetics in the context of transportation asset management. It is important

for decision makers to look across the entire life cycle of an asset. Future research efforts should

include the refinement of LCCA to reflect actual maintenance decisions and realistic

maintenance cycles rather than prescriptive or idealized guidelines. The impact of maintenance

practices on deterioration rates remains an unresolved issue. The American Society of Civil

Engineers Grand Challenge promotes the use of LCCA to extend design life and support

sustainable infrastructure decisions, but it does not provide any specific guidance. There is a

growing need and opportunity for further research in this area.

The topic of adequate subsurface drainage also warrants further research, as the

functionalities of various standard materials are not known quantitatively. Targeted field studies

can provide the required information and fill knowledge gaps, leading to improved subsurface

drainage design guidelines.

IMPLEMENTATION AND BENEFITS

Implementation

With regard to Recommendations 1 through 3, VTRC will work with VDOT’s Materials

Division to help implement changes to the existing geosynthetics specifications and guidelines.

The proposed specification changes are not expected to affect geotextile prices or availability

adversely. According to IFAI’s Specifier’s Guide (IFAI, 2018), more than 70 currently available

nonwoven geotextiles meet the proposed minimum requirements for separation and more than 50

meet the proposed minimum requirements for stabilization. It is anticipated that the Materials

Division will conduct an internal review of proposed recommendations within the next 3 months

and decide on the appropriate outcome with regard to implementation. Meetings with VDOT

district materials engineers will be scheduled shortly to discuss this study. VTRC personnel will

be available for consultation.

With regard to Recommendation 4, VTRC will work with VDOT’s Maintenance

Division to implement changes to the existing PMS. The proposed changes address the need to

incorporate additional data entries required to track the use of geosynthetics. It is anticipated,

based on preliminary discussions, that the Maintenance Division will adopt Recommendation 4

within the next 3 months.

Benefits

The benefits of implementing Recommendations 1 through 3 include the harmonized use

of geosynthetic materials in road construction across the VDOT network in a cost-effective

manner. The revised specifications and guidelines will facilitate consistent, efficient, and

systemic use of separator geotextiles on road projects. It is anticipated that subdivision and

60

secondary roads will benefit appreciably from implementation of these recommendations,

resulting in reduced maintenance costs.

The benefit of implementing Recommendation 4 is that it will allow documentation of

the future use of geosynthetics in roadway construction/rehabilitation. This will provide a

database of geosynthetic pavement sections for future studies.

ACKNOWLEDGMENTS

This study was supported by the Federal Highway Administration. The researchers

express their appreciation to VDOT’s geotechnical and pavement engineers for their technical

guidance. Brian Diefenderfer assisted with ground-penetrating radar testing. Travis Higgs

facilitated field drilling and sampling. Stephen Lane performed the XRD tests. The study would

not have been possible without the extensive field support provided by the maintenance

personnel of VDOT’s Bedford and Charlottesville residencies. Linda Evans of VTRC assisted

with the editorial process.

REFERENCES

Al-Qadi, I.L., and Appea, A.K. Eight-Year Field Performance of Secondary Road Incorporating

Geosynthetics at Subgrade-Base Interface. Transportation Research Record: Journal of

the Transportation Research Board, No. 1849, 2003, pp. 212-220.

Al-Qadi, I.L., Brandon, T.L., Valentine, R.J., Lacina, B.A., and Smith, T.E. Laboratory

Evaluation of Geosynthetic-Reinforced Pavement Sections. Transportation Research

Record: Journal of the Transportation Research Board, No. 1439, 1994, pp. 25-31.

American Association of State Highway and Transportation Officials. AASHTO Guide for

Design of Pavement Structures. Washington, DC, 1993.

American Association of State Highway and Transportation Officials. AASHTO M 288-17:

Standard Specification for Geosynthetic Specification for Highway Applications.

Washington, DC, 2017.

American Association of State Highway and Transportation Officials. National Transportation

Product Evaluation Program. Washington, DC, 2018.

http://www.ntpep.org/Pages/What_is_NTPEP.aspx. Accessed January 2, 2019.

American Railway Engineering Association, Committee on Roadway and Ballast. Roadway and

Ballast. In Proceedings of the Forty-fifth Annual Convention of the American Railway

Engineering Association, Chicago, Vol. 47, 1946, p. 363.

American Railway Engineering and Maintenance-of-Way Association. Manual for Railway

Engineering. Lanham, MD, 2018.

61

Appea, A.K. In-situ Behavior of Geosynthetically Stabilized Flexible Pavement. Master’s

Thesis. Virginia Polytechnic Institute and State University, Blacksburg, 1997.

ASTM International. ASTM D 4439-18: Standard Terminology for Geosynthetics. In Annual

Book of ASTM Standards. West Conshohocken, PA, 2018.

Autostrads Ltd. Guide to Pavement Technology Part 2: Pavement Structural Design. Sydney,

New South Wales, Australia, 2012.

Barksdale, R.D., Brown, S.F., and Chan, F. NCHRP Report 315: Potential Benefits of

Geosynthetics in Flexible Pavement Systems. Transportation Research Board,

Washington, DC, 1989.

Bearden, J., and Labuz, J. Interface Friction of a Soil-Fabric-Aggregate System. In Proceedings

of Cold Regions Impact on Civil Works, American Society of Civil Engineers, Duluth,

Minnesota, 1998, pp. 339-350.

Beckham, W.K., and Mills, W.H. Cotton-Fabric-Reinforced Roads. Engineering News Record,

Vol. 114, No. 14, 1935, pp. 453-455.

Bhutta, S.A. Mechanistic-Empirical Pavement Design Procedure for

Geosynthetically Stabilized Flexible Pavements. Doctoral Dissertation. Virginia

Polytechnic Institute and State University, Blacksburg, 1998.

Black, P.J., and Holtz, R.D. Performance of Geotextile Separators Five Years After Installation.

Journal of Geotechnical and Geoenvironmental Engineering, Vol. 125, No. 5, 1999, pp.

404-412.

Brandon, T.L., Al-Qadi, I.L., Lacina, B.A., and Bhutta, S.A. Construction and Instrumentation

of Geosynthetically Stabilized Secondary Road Test Sections. Transportation Research

Record: Journal of the Transportation Research, No. 1534, 1996, pp. 50-57.

Bruinsma, J., Smith, K., Peshkin, D., Ballou, L., Eisenberg, B., Lurie, C., Costa, M., Ung, C.,

Nassiri, S., Shi, X., and Haselbach, L. ACRP Report 178: Guidance for Usage of

Permeable Pavements at Airports. Transportation Research Board, Washington, DC,

2017.

Christopher, B.R. Cost Savings by Using Geosynthetics in the Construction of Civil Works

Projects. In Proceedings of the 10th International Conference on Geosynthetics (Vol. 1),

September 2014. http://toc.proceedings.com/30777webtoc.pdf. Accessed December 4,

2018.

Christopher, B.R., Schwartz, C., and Boudreau, R. Geotechnical Aspects of Pavements. NHI-

05-037. Federal Highway Administration, Washington, DC, 2006.

Collins, B.M., and Holtz, R.D. Long-Term Performance of Geotextile Separators, Bucoda Test

Site—Phase III. WA-RD 595.1. Washington State Transportation Center, Seattle, 2005.

62

Colorado Department of Transportation. Geosynthetics & Geotextiles. Construction Standard

Specification. Denver, 2005. https://www.codot.gov/business/designsupport/cdot-

construction-specifications/2005-construction-specs/2005ssp/section-200-500-

revisions/208gg.doc/view. Accessed January 2, 2019.

Duong, T.V., Cui, Y.-J., Tang, A.M., Dupla, J.C., Calon, N., Canou, J., and Robinet, A.

Investigating the Mud Pumping and Interlayer Creation Phenomena in Railway Sub-

Structure. Engineering Geology, Vol. 171, 2014, pp. 45-48.

Equipment World. Geosynthetics Speed Construction. May 2014.

https://www.equipmentworld.com/down-to-earth/. Accessed January 2, 2019.

Federal Highway Administration. Life-Cycle Cost Analysis in Pavement Design. September

1998. https://www.fhwa.dot.gov/infrastructure/asstmgmt/013017.pdf. Accessed March

5, 2019.

Federal Register. Discount Rates for Cost-Effectiveness Analysis of Federal Programs. Office

of Management and Budget, The Daily Journal of the United States Government,

February 8, 2018. https://www.federalregister.gov/documents/2018/12/27/2018-

27962/discount-rates-for-cost-effectiveness-analysis-of-federal-programs. Accessed

April 8, 2019.

Gobla, M.J. Design Standards No. 13: Embankment Dams. U.S. Bureau of Reclamation,

Washington, DC, 2014. https://www.usbr.gov/tsc/techreferences/designstandards-

datacollectionguides/finalds-pdfs/DS13-19.pdf. Accessed January 2, 2019.

Greenwood, J.H., Schroeder, H.F., and Voskamp, W. Durability of Geosynthetics. CUR 243.

CUR Building and Infrastructure, 2012. https://www.geosynthetica.net/wp-

content/uploads/Publication243_Durability_C187.pdf. Accessed August 26, 2019.

Hawkins, W. Geotextiles in Unpaved Roads: A 35-Year Case History. July 2009.

http://dbbinc.ca/en/files/TGeo_35yr_Case_History.pdf. Accessed December 5, 2018.

Heukelom, W., and Klomp, A.J.G. Dynamic Testing as a Means of Controlling Pavements

During and After Construction. In Proceedings of First International Conference on the

Structural Design of Asphalt Pavements, Ann Arbor, Michigan, 1962.

Holtz, R. Geosynthetics R&D—The “Early” Days (1960s to Circa 1985). March 9, 2016.

http://www.geosynthetica.com/Uploads/KoernerSympBobHoltz.pdf.

Accessed December 4, 2018.

Holtz, R.D., Christopher, B.R., and Berg, R.R. Geosynthetic Design and Construction

Guidelines. FHWA NHI-07-092. Federal Highway Administration, Washington, DC,

2008.

63

Hossain, S., and Schmidt, B. Benefits of Using Geotextile Between Subgrade Soil and Base

Course Aggregate in Low-Volume Roads in Virginia. VTRC 10-R1. Virginia

Transportation Research Council, Charlottesville, 2009.

Industrial Fabrics Association International. Specifier’s Guide. Roseville, MN, 2018.

https://geosyntheticsmagazine.com/wp-

content/uploads/sites/26/2018/02/Geotextiles_2018.pdf. Accessed January 2, 2019.

Jorenby, B.N., and Hicks, R.G. Base Course Contamination Limits. Transportation Research

Record: Journal of the Transportation Research Board, No. 1095, 1986, pp. 86-101.

Kermani, B., Xiao, M., Stoffels, S.M., and Qiu, T. Reduction of Subgrade Fines Migration Into

Subbase of Flexible Pavement Using Geotextile. Geotextiles and Geomembranes, Vol.

46, 2018, pp. 377-383.

Kermani, B., Xiao, M., Stoffels, S.M., and Qiu, T. Measuring the Migration of Subgrade Fine

Particles Into Subbase Using Scaled Accelerated Flexible Pavement Testing: A

Laboratory Study. Road Materials and Pavement Design, Vol. 20, No. 1, 2019, pp. 36-

57.

Koerner, G.R. Data Base Development for Determination of Long Term Benefit/Cost of

Geotextile Separators. In Proceedings of the Geosynthetics ’97 Conference, Long Beach,

California, 1997, pp. 701-713.

Koerner, R.M. Designing With Geosynthetics, Fifth Edition. Prentice-Hall, Upper Saddle River,

NJ, 2005.

Koerner, R.M. Geosynthetics in Filtration, Drainage and Erosion Control. Elsevier, Essex,

England, 2013.

Koerner, R.M. Designing with Geosynthetics, Sixth Edition. Xlibris LLC, Bloomington, IN,

2016.

Kroening, J.R., Darmiento, F.T., and Hansen, H.E. Geosynthetics—Specifications and

Applications for Arizona. Report SPR-722, Vol. 1. Arizona Department of

Transportation, Phoenix, 2017a. Accessed January 8, 2019.

Kroening, J.R., Darmiento, F.T., and Hansen, H.E. Geosynthetics—Specifications and

Applications for Arizona. Report SPR-722, Vol. 2. Arizona Department of

Transportation, Phoenix, 2017b.

Luo, R., Gu, F., Luo, X., Lytton, R.L., Hajj, E.Y., Siddharthan, R.V., Elfass, S., Piratheepan, M.,

and Pournoman, S. Quantifying the Influence of Geosynthetics on Pavement

Performance. NCHRP Web-only Document 235. Transportation Research Board,

Washington, DC, 2017.

64

Metcalfe, R.C., Holtz, R.D., and Allen, T.M. Field Investigation to Evaluate the Long Term

Separation and Drainage Performance of Geotextile Separators. In Proceeding of the

Geosynthetics 1995 Conference, Vol. 3, Nashville, Tennessee, 1995, pp. 951-962.

Minnesota Department of Transportation. Specification 3733. St. Paul, 2005.

https://www.dot.state.mn.us/products/concrete/pdf/Specification_3733.pdf. Accessed

May 11, 2019.

Missouri Highways and Transportation Commission. Missouri Standard Specifications for

Highway Construction. Jefferson City, 2018.

https://www.modot.org/sites/default/files/documents/2018%20Missouri%20Standard%20

Specific%20-%20MHTC%20%28July%202018%29.pdf. Accessed January 8, 2019.

Montana Department of Transportation. Geotechnical Manual. Helena, 2008.

https://www.mdt.mt.gov/other/webdata/external/materials/geotech_manual/chapter20.pdf.

Accessed May 11, 2019.

Moulton, L. Highway Subdrainage Design. FHWA-TS-80-224. Federal Highway

Administration, Washington, DC, 1980.

Nebraska Department of Roads. Geotechnical Policies and Procedures Manual. Lincoln, 2012.

New York State Department of Transportation. Geotechnical Design Manual: Chapter 18:

Geosynthetic Design. Albany, 2018a.

https://www.dot.ny.gov/divisions/engineering/technical-services/geotechnical-

engineering-bureau/geotech-eng-repository/GDM_Ch-18_Geosynthetic_Design.pdf.

Accessed May 11, 2019.

New York State Department of Transportation. Geosynthetics for Highway Construction.

Albany, 2018b. https://www.dot.ny.gov/divisions/engineering/technical-

services/technical-services-repository/alme/pages/470-1a.html. Accessed May 11, 2019.

Oregon Department of Transportation. Pavement Design Guide. Salem, 2011.

Pennsylvania Department of Transportation. Section 735—Geotextiles. Harrisburg, 2016.

https://www.dot.state.pa.us/public/PubsForms/Publications/Pub_408/408_2016/408_201

6_IE/408_2016_IE.pdf. Accessed January 8, 2019.

Perkins, S.W. Mechanistic-Empirical Modeling and Design Model Development of Geosynthetic

Reinforced Flexible Pavements. FHWA/MT-01-002/99160-1. Federal Highway

Administration, Washington, DC, 2001.

Petrasic, K.W. Geotextile Separation to Prevent Migration of Subgrade Fines. PowerPoint

Presentation. 50th Mid-Atlantic Quality Assurance Workshop, February 7-9, Hershey,

Pennsylvania, 2017.

65

http://www.dot.state.pa.us/public/Bureaus/BOPD/2017_QAW/GeoTech_and_Aggregates

/Geotextile%20Separation-kwp.pdf. Accessed May 11, 2019.

Rankilor, R.R. Membranes in Ground Engineering. John Wiley & Sons Ltd., Chichester,

England, 1981.

Raymond, G.P., and Giroud, J.P. Geosynthetics Case Histories. BiTech Publishers Ltd,

Richmond, British Columbia, Canada, 1993, pp. 192-193.

Robbins, M., and Tran, N. Review of Initial Service Life Determination in Life Cycle Cost

Analysis (LCCA) Procedures and Practice. Report 18-02. National Center for Asphalt

Technology, Auburn, AL, July 2018.

Saathoff, F. Examinations of Long-Term Filtering Behaviour of Geotextiles. In Proceedings of

Durability of Geotextiles Seminar, Saint-Rémy-lès-Chevreuse, France, 1988, pp. 86-114.

Sarsby, R.W. Geosynthetics in Civil Engineering. Woodhead Publishing Limited, Cambridge,

England, 2006.

Shukla, S.K. An Introduction to Geosynthetic Engineering. Taylor & Francis Group, London,

2016.

Skok, E.I., Johnson, E.N., and Brown, M. Special Practices for Design and Construction of

Subgrades in Poor, Wet and/or Saturated Soil Conditions. Minnesota Department of

Transportation, St. Paul, 2003.

Suits, L.D., and Richardson, G.N. M288-96: The Updated AASHTO Geotextile Specifications.

Geotechnical Fabrics Report, Vol. 16, No. 2, 1998, pp. 39-41.

Ullah, S., Tanyu, B.F., Guler, E.F., Hoppe, E.J., and Akmaz, E. Evaluation of the Long-Term

Performance of Woven Geotextile Used Between Base Course and Subgrade of a Paved

Road. Transportation Research Record: Journal of the Transportation Research Board,

No. 2673, 2019, pp. 310-321.

U.S. Department of Defense. Engineering Use of Geotextiles. UFC 3 220 08fa 2004.

Washington, DC, 2004.

https://www.wbdg.org/FFC/DOD/UFC/ufc_3_220_08fa_2004.pdf. Accessed October 8,

2019.

U.S. Office of Management and Budget. Discount Rates for Cost-Effectiveness, Lease Purchase,

and Related Analyses. OMB Circular No. A-94. https://www.whitehouse.gov/wp-

content/uploads/2018/12/Appendix-C.pdf. Accessed March 5, 2019.

Van Santvoort, G. Geotextiles and Geomembranes in Civil Engineering. A. A. Balkema,

Rotterdam, The Netherlands, 1994.

66

Virginia Department of Environmental Quality. Permeable Pavement. Richmond, 2013.

https://www.swbmp.vwrrc.vt.edu/wp-

content/uploads/BMP_Spec_No_7_PERMEABLE_PAVEMENT.pdf. Accessed June 18,

2019.

Virginia Department of Transportation. Flexible Pavement Design Guide for Primary and

Interstate Roads in Virginia. Richmond, 1995.

Virginia Department of Transportation. Pavement Design Guide for Subdivision and Secondary

Roads in Virginia. Richmond, 2000.

Virginia Department of Transportation. Road Design Manual, Chapter 2E, Section 2E-1.

Richmond, 2005 [Rev. January 2019].

http://www.virginiadot.org/business/resources/LocDes/RDM/chap2e.pdf. Accessed June

18, 2019.

Virginia Department of Transportation. 2013 Lump Sum Design-Build Agreement Between

Department and Design-Builder. Richmond, 2013.

http://www.virginiadot.org/business/resources/2013_Parts_3-4-5_Documents_7-3-

2013.pdf. Accessed June 18, 2019.

Virginia Department of Transportation. 2016 Road and Bridge Specifications. Richmond, 2016.

Virginia Department of Transportation. Pavement Design Guide for Subdivision and Secondary

Roads in Virginia. Richmond, 2018a.

Virginia Department of Transportation. Chapter 6—Pavement Design and Evaluation. In

Manual of Instructions, Materials Division, Virginia Department of Transportation,

Richmond, August 2018b. http://www.virginiadot.org/business/resources/Materials/bu-

mat-MOI-VI.pdf. Accessed January 2, 2019.

Virginia Department of Transportation. State of the Pavement Reports. Richmond, 2018c.

http://www.virginiadot.org/info/state_of_the_pavement.asp. Accessed January 8, 2019.

Virginia Department of Transportation. Approved List No. 63: Geosynthetics (Soils Lab).

Richmond, 2018d.

Virginia Department of Transportation. Design Build Template Request for Proposal.

Richmond, 2019.

Washington State Department of Transportation. WSDOT Design Manual: Chapter 630:

Geosynthetics. M 22-01.10. Olympia, 2016.

http://www.wsdot.wa.gov/publications/manuals/fulltext/M22-01/630.pdf. Accessed

December 10, 2018.

67

Whitaker, B.A. 40-Year Study of Geotextiles in an Unpaved Road. Transportation Research

Board Annual Conference, 2014. https://trid.trb.org/view/1288672. Accessed October 4,

2019.

Woods, K.B., and Shelburne, T.E. Pumping of Rigid Pavements in Indiana. In Proceedings of

the Twenty-third Annual Meeting of the Highway Research Board. Highway Research

Board, Washington, DC, 1943, pp. 301-316.

Xiao, M., Stoffels, S., Qui, T., and Kermani, B. Evaluation of Geotextile Separation to Prevent

Migration of Subgrade Fines into Subbase. Pennsylvania State University, State College,

2016.

Yang, S.-H., and Al-Qadi, I.L. Cost-Effectiveness of Using Geotextiles in Flexible Pavements.

Geosynthetics International, Vol. 14, No. 1, 2007, pp. 2-12.

Zornberg, J.G. Functions and Applications of Geosynthetics in Roadways. Procedia

Engineering, Vol. 189, 2017, pp. 298-306.

Zornberg, J.G., and Thompson, N. Application Guide and Specifications for Geotextiles in

Roadway Applications. FHWA/TX-10/0-5812-1. Texas Department of Transportation,

Austin, 2012.

68

69

APPENDIX A: SURVEY OF VDOT RESIDENT ENGINEERS

1. In general, is it your experience that subdivision streets that are accepted into VDOT's

secondary system reach their 20-year design life?

No 71.4%

Yes 21.4%

Usually yes 7.1%

2. If not, in your opinion are their deficiencies more related to pavement design factors,

construction quality, or an inaccurate forecast of traffic loading?

Pavement design factors 40.0%

Construction quality 46.7%

All, esp. traffic forecast 6.7%

No answer 6.7%

3. In your opinion, are subdivision streets that are accepted into VDOT’s secondary system

of the same level of quality (i.e. as durable) as VDOT secondary roadway projects?

Yes 33.3%

No 66.7%

4. If not, please describe quality differences you have noted.

Surface issues with asphalt;

Unsuitable subgrade material and installation, differential settlement, insufficient pavement

foundation drainage, underdrains, crossdrains, etc.;

Pavement thickness over the length of road is inconsistent and subgrades with unsuitable

material can be overlooked;

Less field oversight during construction of subdivision streets as compared to secondary

roadway projects with dedicated inspection force;

VDOT oversight during secondary roadway construction is far greater than subdivision streets

due to staff and the way a project is facilitated compared to developer built streets. VDOT does

have an inspection manual for subdivision streets, but currently my Residency has 1 inspector

for 4 Counties;

VDOT secondary road specs have a more robust design. Most VDOT roadway projects meet or

exceed the designed life;

VDOT's minimum pavement specifications for subdivision streets are not sufficient;

Secondary road QA/QC programs are lacking. Pavement designs are insufficient and allowable

phasing is set up to allow the developer to minimize exposure during his maintenance period

and maximize traffic load after their maintenance period;

Lack of inspection allows substandard construction. Pipe bedding is poor, curb and gutter and

sidewalk construction is substandard;

Pavement failures occur much earlier than VDOT roadways;

Drainage is probably the biggest problem, erosion of ditch lines and undersized cross and

entrance pipes. Depth of aggregate.

5. In your experience, what type of maintenance performed on subdivision streets has the

highest cost?

Pavement rehabilitation

70

Asphalt resurfacing

Premature overlay, deep culvert failure/replacement

Repaving, sidewalk reconstruction, closed drainage failures

Mill/Overlay is a normal method however FDR is becoming popular, is expensive but corrects

base concerns that ultimately extends service life and condition.

Drainage infrastructure. Replacing pipes, sinkholes, sliplining, etc.

Heavy mechanized patching

Patching and filling potholes.

Resurfacing

Pavement

Pavement settlements and pavement failures

Pavement overlay

Patching and overlay

Pavement and drainage work

Resurfacing of various types

6. Approximately what percentage of your maintenance budget for asphalt paving is spent

on subdivision streets?

Low. Our percentage of subdivision streets are low, and we can use low cost options such as

slurry seal.

15%

Increasing - I will email specific percentage

10%

5% because I do not have as many subdivisions in our rural location.

1%

--

In NOVA, the Infrastructure Office controls the paving budget and not the Residency

Maintenance Administrator.

40%

--

Unsure but the funding available is insufficient to keep up with the needs.

< 5%

Surface treatment and slurry on subdivision streets are approximately 20%. We do not overlay

with asphalt.

15% to 20%

10%

7. For subdivision streets that have exceeded their design life (i.e. require reconstruction),

what method is typically used to reconstruct these streets?

If repairs beyond slurry seal are required, mill and repave due to curb and gutter sections on

many of the streets.

Hired equipment/Maintenance Forces

Surface-treatment unless County/subdivision contribute funds to use another resurfacing

solution

71

Mill and overlay surface asphalt. I can't think of a time we have done a complete rebuild to a

subdivision street outside of spot areas.

Base repair, patch and then overlay. FDR is being looked into as an option for streets with wide

spread base failure.

Slurry or mill/replace.

When they are resurfaced a base layer of asphalt may be added. Full road reconstruction has

such a large cost associated with it that it is only used in very rare circumstances.

If it's full mill then it's paved in kind. At times they do a curb mill with a layover.

Mill and overlay and possibly FDR.

--

Very rare we have the funds to reconstruction but when we do we will typically remove the

existing surface undercut in failing areas and replace with a double or triple shot of ST.

Rarely totally reconstruct subdivision streets. Some milling and patching may occur.

Scarify and resurface.

Overlays, dig up soft spots, and upgrading pipes.

Spot leveling, then repaving.

8. Do the majority of your subdivision streets include some form of subsurface drainage

system (i.e. underdrains, cross-drains, daylighted aggregate, etc.)?

Yes 40.0%

No 60.0%

9. Is contamination of the aggregate layer by the subgrade soil a cause of pavement

degradation in your residency?

Always 0%

Sometimes 92.9%

Never 0%

Sometimes depending on the

location 7.1%

10. Do you know of any examples in your residency of the use of geotextile to separate the

soil subgrade and aggregate layers?

Yes 26.7%

No 73.3%

Comments:

No, but we feel that would be a good practice.

Yes, the Research Council installed a couple areas on Rt. 743 in Albemarle around the airport

back around 2008

Yes. The majority of the time this is used in spot locations during construction when

undercutting is not feasible due to conditions.

Yes, this method has been used in various developments

No, not in subdivisions. We need more aggregate under these roads.

72

73

APPENDIX B: LAYER DETERIORATION CURVES AND MAINTENANCE

SCHEDULE DETERMINATION PROCESS SUPPORTING THE LCCA

Each initial pavement design compared in the LCCA is consistent with the inputs given

in Table 1 in the body of the report and with AASHTO (1993) (per VDOT Materials Division

Manual of Instructions, Chapter 6) in order to determine the remaining design life in all analysis

years of the LCCA.

As discussed in the body of the report, the LCCA model developed in this study depends

on proposed deterioration curves for secondary road pavement layers with and without separator

geotextile. Although based on the AASHTO approach, the proposed deterioration curves reflect

the consensus mechanism by which pavement structure deterioration occurs because of subbase

contamination with subgrade fines:

Aggregate subbase/base contamination: with and without separator geotextile,

proposed to model aggregate contamination by subgrade soils when a separator is not

used. The contamination rate was varied over a range including 0.2 in/yr in the base

case, and 0.1 in/yr and 0.05 in/yr, respectively, in lower contamination scenarios.

Asphalt base mixture: with and without separator geotextile, to reflect stripping and

accelerated deterioration because of loss of aggregate subbase.

Asphalt surface mixture: over both competent and poor asphalt base mixtures to

simulate overlays over cracked existing pavements where existing cracks propagate

through the new surface mixture. A maximum service life of surface asphalt was

conservatively set at 20 years to reflect typical mixture serviceability and reduced

maintenance on the secondary roadway system.

The AASHTO Guide for Design of Pavement Structures (AASHTO, 1993) provides the

basis for the deterioration curves proposed for this analysis. Figure B1 shows the widely

accepted serviceability and structural capacity curves; Figure B2 illustrates condition factors

based on the best-fit curve from the AASHO road tests.

It is the predominant understanding, however, that bound pavement materials do not

deteriorate suddenly to the structural coefficient of zero. Thus, the modified “idealized”

deterioration curves shown in Figure B3 were developed to represent engineering experience and

the consensus regarding the function of separator geotextiles. Aggregate contamination can be

idealized in two different ways: the aggregate layer thickness remains the same, but the structural

coefficient of the aggregate layer decreases over time or the structural coefficient of the

aggregate layer remains the same but the thickness of the aggregate layer diminishes with time.

The latter method was selected by the researchers to allow for more clarity in pavement design

calculations.

74

Figure B1. Serviceability and Structural Capacity (AASHTO, 1993)

Figure B2. Condition Factor vs. Remaining Life Curves (AASHTO, 1993)

75

Figure B3. Idealized Deterioration Curves With and Without Separator Geotextile Under Two

Contamination Rates: 1 in in 5 years and 1 in in 10 years

0

1

2

3

4

5

6

0 5 10 15 20 25 30 35 40 45 50

Rem

ain

ing

Laye

r (i

n.)

Years

Geotextile

No Geotextile (1 in/5yrs)No Geotextile (1in/10 yrs)

(b) Aggregate Contamination vs. Remaining Layer Depth

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 5 10 15 20 25

Stru

ctu

ral C

oef

fici

ent

Years

Surface Mix over BM > 0.33

Surface Mix over BM < 0.33

1993 AASHTO Best Fit Curve

Limit of Serviceable Surface Mix

(d) Surface Mixture Deterioration

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0 5 10 15 20 25 30 35 40 45 50

Stru

ctu

ral C

oef

fici

ent

Years

Geotextile

No Geotextile (1in/5 yrs)No Geotextile (1in/10 yrs)

(a) Aggregate Contamination vs. Structural Coefficient

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

0 5 10 15 20 25

Stru

ctu

ral

Nu

mb

er (

Cx

* SN

o)

Years(e) AASHTO Asphalt Pavement Deterioration

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 5 10 15 20 25 30 35 40 45 50

Stru

ctu

ral C

oef

fici

ent

Years

Geotextile

No Geotextile (1 in/5 yrs)

No Geotextile (1 in/10 yrs)

Limit of Serviceable Base Mix

(c) Base Mixture Deterioration

76

In two recent reports, a comprehensive contract study for PennDOT (Xiao et al., 2016)

and a streamlined derivative report in Kermani et al. (2019), accelerated pavement testing

equipment (i.e., MMLS3, or model mobile load simulator at 1/3 scale) was used to determine

quantified rates of subgrade fines penetration into aggregate subbase. Some selected parameters

of the PennDOT study are shown in Table B1 in comparison with typical VDOT values.

The results from the PennDOT study indicate that fines contamination of the aggregate

subbase layer begins virtually immediately and increases with increasing “cycles,” which for the

modelled collector road section of both studies was essentially equivalent to ESALs (Kermani et

al. 2019; Xiao et al. 2016). The studies examined fines contamination of the top and bottom

aggregate strata as well as full-depth contamination.

A previous study demonstrated that, starting with aggregate already containing 5.5%

fines, the resilient modulus of a contaminated aggregate subbase drops precipitously at

approximately 13.5% fines content (Jorenby and Hicks, 1986). Given VDOT No. 21B gradation

and production tolerances shown in Table B1, additional fines content of 7% was identified as a

conservative contamination level for a potentially abrupt loss of resilient modulus in VDOT

aggregate No. 21B. (Note: Table B1 indicates that VDOT No. 21A aggregate can reach 13.5%

fines even without exceeding production tolerances for No. 21A, and thus the level of

contamination of No. 21A aggregate that could suddenly impair resilient modulus is much lower

than the corresponding one for No. 21B.)

Based on the best linear fit relationship between pavement loadings and full-depth

laboratory contamination findings in Kermani et al. (2019), the laboratory findings are (1)

extended to higher contamination rates within the Kermani model and (2) scaled from the

modeled collector facility design of Kermani et al. (2019) to the two low-volume pavement

designs in the LCCA, as described in Table B1. Figure B4 shows the resulting plots for ESAL-

to-percent of full-depth contamination for the LCCA pavement designs along with Kermani’s

extended findings.

Table B1. Comparison of PennDOT Study Parameters to Typical VDOT Values

Parameter PennDOT Study VDOT

Subgrade Soil

Soil Type A-4 (ML) A-4, A-6, A-7-5, A-7-6

ML, CL, MH, CH

Fines Content 55.9% 35 – 100%

Density AASHTO T 180 Proctor AASHTO T 99 Proctor

Soaked CBR 5 (initial) 3 to 8

Saturation Inundated Variable

Aggregate (VDOT Production Tolerance)

Fines Content 6.5% 4 – 7% (2-9%) (No. 21B)

6 – 12% (4-14%) (No. 21A)

Max Aggregate Size --- 1 in

Pavement Structure

Aggregate Subbase 6 in 8 in (155 AADT)

5 in (1500 AADT)

Asphalt 8.5 in 1.5 in (155 AADT)

6.0 in (1500 AADT)

ML = silt; CL= lean clay; MH = elastic silt; CH = fat clay; AADT = annual average daily traffic.

77

Figure B4. Collector Roadway Loadings, LCCA Secondary Road Loadings, and Resulting Full-Depth

Contamination Rates. Source: Kermani et al. (2019).

The combined findings of Kermani (2019) and Jorenby and Hicks (1986) provide a check

on the range of absolute contamination rates assumed in this study. At the 155 AADT (initial)

pavement (growing at 2% per year), the 7% fines contamination threshold for the full aggregate

depth (82,000 ESALs) is back-calculated by means of AASHTO (1993) to occur at

approximately 94 years after construction. At the 1,500 AADT pavement (also growing at 2%

per year), the 7% threshold is reached at about 327,000 ESALs, or about 18 years after

construction. The corresponding absolute annual rates of contamination are calculated at 0.064

in and 0.341 in, respectively.

The absolute contamination rate parameters used in the LCCA model lie within that

spread: a minimum of 0.05 in per year, 0.1 in per year, and a maximum of 0.2 in per year. These

rates are consistent with the properties of VDOT aggregate and typical VDOT pavement

structures shown in Table B1, as evidenced by the calculation of the contamination rate interval

(0.064, 0.341) based on VDOT values in Table B1. Yet it is important to recognize that the

threshold contamination percentages of fines indicate the point at which the aggregate layer loses

resilient modulus rather than the time when contamination initiates. In fact, without a GTX

separator, the aggregate subbase layer is subjected to contamination virtually concurrently with

the initial traffic loading. Below the threshold value of 7%, the aggregate is contaminated to

some degree; above the 7% threshold, the aggregate has probably lost nearly all of its resilient

modulus and is exhibiting behavior similar to that of the underlying subgrade soil.

The ESAL values shown in Figure B4 are technically ESALs applied while the subgrade

soil is saturated. In other words, if the pavement structure is designed and constructed with an

adequate subsurface drainage system and the subgrade is not allowed to approach full saturation,

then virtually no contamination can occur. Conversely, without separator GTX and assuming

full saturation, subgrade soils will contaminate the aggregate layer, in keeping with the findings

of Kermani et al. and Jorenby and Hicks. Effectively, it means that if the subgrade is saturated

only one-half of its life, the actual contamination rate will be approximately one-half of those

0

1

2

3

4

5

6

7

8

9

0 100 200 300 400 500 600

Co

nta

min

atio

n r

ate

(%)

Equivalent Single Axle Loads (Thousands)

Kermani ESALs

155 AADT

1500 AADT

78

shown. The actual saturation rate over a pavement’s service life, however, is typically

unquantifiable at the time of construction.

Significantly, when Kermani’s contamination rate curve for interstates is applied to a

typical interstate section of I-81 in Virginia (24,000 AADT, 28% tractor trailers, and 3% single

unit trucks), it indicates a contamination rate of approximately 2.47 in/yr. This is a very high

rate of contamination, especially considering the very thick pavement section (17.5 in of

asphalt), but not inconceivable, as shown in Figure 5 in the body of the report.

The remaining design life of each pavement structure is determined for each 5-year

period using the 1993 AASHTO pavement design method (AASHTO, 1993). Figure B5

illustrates the following steps for each calculation:

1. Step 1: Increase traffic data to reflect traffic at time i.

2. Step 2: Reduce structural coefficients of asphalt layers to reflect deterioration at time

i.

3. Step 3: Reduce aggregate layer (if applicable) to reflect contamination at time i.

4. Step 4: Iteratively change Performance Period until Calculated Design Structural

Number (required) equals the Total Calculated Structural Number (of the pavement

structure).

5. Step 5: The Performance Period at which these structural numbers is equal is taken to

be the remaining life of this pavement structure. This value is captured in the

maintenance history for this alternative.

Maintenance activities are determined based on the following decision criteria:

Surface condition: When the structural coefficient of the surface mixture falls below

0.30, overlay or mill/resurfacing is required. Preventive maintenance was modeled as

milling of the existing asphalt surface mixture and inlay with new asphalt surface

mixture. “No Preventive Maintenance” was modeled as simply overlay with new

asphalt surface mixture. In this analysis, the service period of initial surface asphalt

mixtures was taken to be 20 years to reflect typical reduced maintenance on the

secondary roadway system.

Remaining design life: When the remaining design life falls below 5 years (the

duration between the current and next analysis year), some amount of patching is

required. Although not directly known at the time of the maintenance activity in the

field, this is typically intuited through the necessity for patching. No patching is

performed when resurfacing or rebuilding is performed in the same analysis year.

79

Figure B5. Remaining Life Calculation Example

Rehabilitation/rebuild: If the typical maintenance activity (overlay or mill/inlay) does

not result in at least 5 years of remaining design life, rehabilitation or rebuild is

necessary. This pavement design calculation includes engineering judgment,

although the designs in this analysis minimally meet a new design life of 20 years.

However, the first maintenance activity performed consists of overlay or mill/inlay

rather than rehabilitation/rebuild, in accordance with typical maintenance procedures.

These decision criteria are evaluated in 5-year increments for each pavement alternative

for the 50-year analysis period. The maintenance histories determine the pay items input into the

80

LCCA. Some maintenance activities are not intuitive, and some are estimated, including the

following:

Twenty percent of the roadway length (both sides) requires guardrail.

Guardrail replacement is necessary after overlay exceeds 2 in.

Crack sealing is considered a routine maintenance activity common to all pavements

and is therefore not included in the pavement maintenance histories.

81

APPENDIX C: EXAMPLE OF MAINTENANCE SCHEDULE DETERMINATION

PROCESS: NO GTX PAVEMENT OPTION WITH AND WITHOUT PM

The management of pavement options for a 50-year life-cycle cost analysis period is

more complex than the standard application of maintenance rules at given points in time.

Pavement management for the LCCA combines the decision criteria shown in Figure 3 in the

body of the report with engineering judgment to anticipate actual maintenance decisions under

the real-world constraint of cost-effective asset management. To summarize this hybrid process,

in each analysis year, the surface structural coefficient determines whether an activity is

warranted. If it is not warranted and the existing remaining design life is under 5 years, routine

patching is performed. If it is warranted, the actual maintenance activities common to low-

volume roads are then evaluated using engineering judgment, starting with lowest expense

activities. If the recalculated remaining design life does not exceed 5 years, the maintenance

activity is upgraded to rehabilitation; if recalculated and the remaining design life still remains

under 5 years, maintenance is upgraded further to rebuilding of the road.

Figure C1 summarizes the final maintenance schedule for the No GTX-No PM pavement

option over the 50-year analysis period. The decision process proceeds as follows: by 2040 the

surface structural coefficient is below 0.3 and patching is insufficient to provide more than 5

years of service life. Overlay is therefore selected, providing 11.4 years of service life. Fifteen

years later, however, patching is no longer sufficient and even another overlay gives only slightly

more than 5 years of service life. In 2065, the overlay provides less than 5 years of service life

and rehabilitation results in less than 20 years of service life. Reconstruction is elected at that

point to provide 20 additional years of service life. The resulting 50-year maintenance schedule,

and the costs corresponding to this pavement option, is presented in detail in Appendix D, Figure

D1.

Figure C1. Example of Management of Pavement Without Geotextile or Preventive Maintenance. ERDL =

estimated remaining design life; SSC = surface structural coefficient; AADT = annual average daily traffic;

RRDL = recalculated remaining design life; SM = surface mix.

Year ERDL Maintenance SSC AADT

2020 20 Construction 0.440 155

2030 6.2 0.395 189

2035 2.8 Patch 0.358 209

2040 0.8 Overlay 1.5" (RRDL = 11.4 yrs) 0.266 230

2045 4.3 Patch 0.421 254

2050 1.3 Patch 0.395 281

2055 0.31 Overlay 1.5" (RRDL = 5.6 yrs) 0.358 310

2060 2.3 Patch 0.421 342

2065 0.51 Rebuild 8.5" 21B, 2" SM (RRDL = 20.0 yrs) 0.395 378

2070 12.3 Remaining service life = 12.3 yrs 0.421 417

No GTX-No PM, Initial AADT 155, Contamination Rate 0.2-in/yr

82

Figure C2 summarizes the final maintenance schedule for the No GTX-PM pavement

option over the 50-year analysis period. The reasoning in Figure 3 in the body of the report

governs maintenance decisions made for this pavement scenario, as it does for the scenario

shown in Figure C1. The 50-year maintenance schedule for this pavement scenario is shown in

detail in Figure D2 in Appendix D. Figures D1 and D2 show that Preventive Maintenance in the

LCCA model includes milling of deteriorated pavement whereas pavements without preventive

maintenance are overlaid without surface milling.

Figure C2. Example of Management of Pavement Without Geotextile But With Preventive Maintenance.

ERDL = estimated remaining design life; SSC = surface structural coefficient; AADT = annual average daily

traffic; RRDL = recalculated remaining design life; SM = surface mix.

Year ERDL Maintenance SSC AADT

2020 20.0 Construction 0.440 155

2030 6.2 0.395 189

2035 2.8 Patch 0.358 209

2040 0.80 Mill/Inlay 1.5" (RRDL = 2.6 yrs) 0.266 230

2045 1.2 Patch 0.421 254

2050 0.51 Rebuild 8" 21B, 2" SM (RRDL = 22.5 yrs) 0.395 281

2055 13.4 0.421 310

2060 7.2 0.395 342

2065 3.3 Patch 0.358 378

2070 0.82 Mill/Inlay 2" (RRDL = 3.5 yrs) 0.266 417

No GTX-PM, Initial AADT 155, Contamination Rate 0.2-in/yr

83

APPENDIX D: EXAMPLES OF PAVEMENT MAINTENANCE SCHEDULES WITH

AND WITHOUT PREVENTIVE MAINTENANCE

Figure D1. Maintenance Schedule for No GTX-No PM Pavement Option, Initial AADT 155, Contamination

Rate 0.2 in/yr. AADT = annual average daily traffic.

Mainline - HMA Surface 1.5 1,156 Tons 93$ 107,052$ 107,052$

Mainline - 21B 8 6,420 Tons 23$ 146,515$ 146,515$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 10,000$

263,567$ 263,567$ CEI (15%) 39,535$ 39,535$

303,103$ 303,103$

Mainline - Patching 44% 1.5 509 Tons 350$ 178,073$ 98,878$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 5,553$

188,073$ 104,430$

CEI (5%) 9,404$ 5,222$

197,477$ 109,652$

Mainline - Overlay HMA Surface1.5 1,156 Tons 93$ 107,052$ 48,857$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 4,564$

117,052$ 53,421$

CEI (5%) 5,853$ 2,671$

122,905$ 56,092$

Mainline - Patching 14% 1.5 162 Tons 350$ 56,660$ 21,254$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 3,751$

66,660$ 25,005$

CEI (5%) 3,333$ 1,250$

69,993$ 26,255$

Mainline - Patching 74% 1.5 856 Tons 350$ 299,487$ 92,337$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 3,083$

309,487$ 95,421$

CEI (5%) 15,474$ 4,771$

324,961$ 100,192$

Mainline - Overlay HMA Surface1.5 1,156 Tons 93$ 107,052$ 27,129$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 2,534$

117,052$ 29,663$

CEI (5%) 5,853$ 1,483$

122,905$ 31,146$

Mainline - Patching 54% 1.5 624 Tons 350$ 218,544$ 45,520$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 2,083$

228,544$ 47,603$

CEI (5%) 11,427$ 2,380$

239,972$ 49,983$

Demolition of Pavement 14,080 SY 12$ 168,960$ 28,926$

Mainline - HMA Surface 2 1,542 Tons 91$ 139,529$ 23,887$

Mainline - 21B 8.5 6,822 Tons 23$ 154,786$ 26,499$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 1,712$

473,275$ 81,024$

CEI (15%) 70,991$ 12,154$

544,266$ 93,178$

12.3 years PV of 50-year Cost 769,601$ Remaining Useful Life207050

205535

206040

206545

204020

204525

205030

Unit Cost Total Present Value

20200

Depth

(in)Quantity Unit

203515

Analysis

Year

Calendar

YearActivity

84

Figure D2. Maintenance Schedule for No GTX-PM Pavement Option, Initial AADT 155, Contamination

Rate 0.2 in/yr. AADT = annual average daily traffic.

Mainline - HMA Surface 1.5 1,156$ Tons 93$ 107,052$ 107,052$

Mainline - 21B 8 6,420$ Tons 23$ 146,515$ 146,515$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 10,000$

263,567$ 263,567$

CEI (15%) 39,535$ 39,535$

303,103$ 303,103$

Mainline - Patching 44% 1.5 509$ Tons 350$ 178,073$ 98,878$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 5,553$

188,073$ 104,430$

CEI (5%) 9,404$ 5,222$

197,477$ 109,652$

Mainline - Mill SM Layer 1.5 14,080$ SY 3$ 37,734$ 17,221$

Mainline - Overlay HMA Surface1.5 1,156$ Tons 93$ 107,052$ 48,857$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 4,564$

154,787$ 70,643$

CEI (5%) 7,739$ 3,532$

162,526$ 74,175$

Mainline - Patching 76% 1.5 879$ Tons 350$ 307,581$ 115,379$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 3,751$

317,581$ 119,130$

CEI (5%) 15,879$ 5,957$

333,460$ 125,087$

Demolition of Pavement 14,080$ SY 12$ 168,960$ 52,094$

Mainline - HMA Surface 2 1,542$ Tons 91$ 139,529$ 43,019$

Mainline - 21B 8 6,420$ Tons 23$ 146,515$ 45,173$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 3,083$

465,005$ 143,370$

CEI (15%) 69,751$ 21,505$

534,755$ 164,875$

Mainline - Patching 34% 1.5 393$ Tons 350$ 137,602$ 23,557$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 1,712$

147,602$ 25,269$

CEI (5%) 7,380$ 1,263$

154,982$ 26,533$

Mainline - Mill SM Layer 2 14,080$ SY 3$ 37,734$ 5,310$

Mainline - Overlay HMA Surface2 1,542$ Tons 91$ 139,529$ 19,634$

Maintenance of Traffic 1.00 LS 10,000$ 10,000$ 1,407$

187,264$ 26,350$

CEI (5%) 9,363$ 1,318$

196,627$ 27,668$

3.5 years PV of 50-year Cost 831,091$

206545

207050

RemainIng Useful Life207050

204020

204525

205030

203515

Present Value Total Unit Cost UnitQuantityDepth

(in)Activity

Calendar

Year

Analysis

Year

20200

85

APPENDIX E: PROPOSED SPECIAL PROVISION FOR SUBGRADE SEPARATION

AND SUBGRADE STABILIZATION GEOTEXTILES

VIRGINIA DEPARTMENT OF TRANSPORTATION

SPECIAL PROVISION FOR

SUBGRADE SEPARATION AND SUBGRADE STABILIZATION GEOTEXTILES

October 7, 2019

SECTION 245.03(d) – Geotextile for Use in Stabilization of the Specifications is replaced with

the following:

Geotextile for Use in Separation and Stabilization: These areSeparation geotextiles

are used as a permeable layer to separate fine-grained subgrades and aggregate base or

subbase. Stabilization geotextiles are used in saturated and/or unstable conditions to

provide the functions of separation and reinforcement.

1. Subgrade SeparationStabilization Fabric:

Physical Property Test Method Requirements

Apparent opening size ASTM D 4751 Max. No. 7020 (0.212 mm)

sieve

Permittivity ASTM D4491 Min. 0.1 sec-1

In addition to this requirement, tThe geotextile shall comply with the requirements of

AASHTO M 288 for strength property requirements, Table 1, Class 32, for grab

strength, tear strength, and puncture strength. Only nonwoven geotextiles shall be

used as subgrade separation geotextiles.

2. Subgrade Stabilization Fabric:

Physical Property Test Method Requirements

Apparent opening size ASTM D4751 Max. No. 70 (0.212 mm)

sieve

Permittivity ASTM D4491 Min. 0.1 sec-1

The geotextile shall comply with the requirements of AASHTO M 288 for strength

property requirements, Table 1, Class 1, for grab strength, tear strength, and

puncture strength. Geotextiles used for subgrade stabilization shall be woven or

nonwoven.

86

3. Embankment Stabilization Fabric Up to 6 Feet High:

Physical Property Test Method Requirements

Apparent opening size ASTM D 4751 Max. No. 20 sieve

Seam strength ASTM D 4632 90% specified grab

strength

In addition to this requirement, tThe geotextile shall comply with the requirements

of AASHTO M288 for strength property requirements, Table 1, Class 1 for grab

strength, tear strength, and puncture strength.

SECTION 305.03(d) – Geotextile (Subgrade Stabilization) of the Specifications is replaced

with the following:

Geosynthetics: Geosynthetics include Geotextile (Subgrade Separation) and Geotextile

(Subgrade Stabilization) in accordance with Section 245, and geogrid.

1. Subsurface preparation: Before placing the geotextile, geogrid, or

combination of both, prepare the subgrade in accordance with Sections 304

and 305. Separation and stabilization geotextiles shall not be placed when

weather conditions, in the opinion of the Engineer, are not suitable to allow

placement of geotextiles or cover materials. These include wet or snowy

conditions, rainfall, temperatures below freezing, frost, extreme heat, or

excessively windy conditions.

2. Geotextile or geogrid placement. Place geogrid on top of geotextile when

both are shown at the same elevation in the plans. Place the geosynthetic in

the direction of traffic. Geosynthetic shall be smooth and free of wrinkles and

folds. Placement by dragging the geosynthetic across the finished surface will

not be allowed. On curves, the geotextile may be folded or cut to conform to

the curve. The fold or overlap shall be in the direction of traffic and held in

place by pins, staples or piles of aggregate subbase or base materials.

Overlap in the direction of construction. Overlap at least 24 inches at the ends

and sides of adjoining sheets or sew the joints according to the

Manufacturer’s recommendations. Do not place longitudinal overlaps below

anticipated wheel loads or joints. Hold the geosynthetic in place with pins,

staples, or piles of aggregate subbase or base materials.

Replace or repair geosynthetic that is torn or punctured. Remove the damaged

area and place a patch of the same type of geosynthetic overlapping 36 inches

beyond the damaged area or sew a seam around the entire perimeter of the

damaged area.

3. Initial layer placement and compaction: Place initial layer in accordance

with Sections 308 and 309.

87

Tracked equipment shall not be operated directly on top of geosynthetic.

Rubber-tire equipment may pass over the geosynthetic at lower speed without

causing excessive rutting (2 inches or greater) or causing damage to the

geosynthetic. Contractor shall avoid sudden braking or sharp turning.

If equipment causes subgrade rutting in excess of 2 inches, end dump the

backfill material onto the geotextile or geogrid from the edge of the

geosynthetic or from previously placed cover material. Do not operate

equipment directly on the geosynthetic. Spread the end-dumped pile of cover

material maintaining the minimum specified lift thickness over the

geosynthetic. Avoid sudden stops, starts, or turns of the construction

equipment. Fill ruts from construction equipment with additional cover

material. Do not blade material down to remove ruts. If rutting exceeds 2

inches during placement, decrease the construction equipment size, decrease

the equipment weight, or increase the first lift thickness as directed by the

Engineer.

Compact in accordance with Sections 308 and 309. Do not use sheepsfoot or

studded compaction equipment. Compact the cover material with pneumatic-

tire or non-vibratory smooth drum rollers.

4. Subsequent layer placement and compaction. Place and compact subsequent

layers in accordance with Sections 308 and 309.

(d) Geotextile (Subgrade Stabilization) : When geotextile for subgrade stabilization is

required, it shall be placed as shown on the plans. Geotextile shall be spliced by an overlap

of at least 2 feet or by sewing double-stitched seams with stitching spaced ¼ inch to ½ inch

apart or as shown on the plans.

SECTION 305.04 – Measurement and Payment of the Specifications is amended to add the

following:

Geotextile for subgrade stabilization will be measured in square yards, complete-in-

place. Overlaps and seams will not be measured for separate payment. The accepted

quantity of geotextile will be paid for at the contract unit price per square yard price. This

price shall include furnishing, placing, lapping, or seaming material.

Payment will be made under:

Pay Item Pay Unit

Geotextile (Subgrade Separation) SY

Geotextile (Subgrade Stabilization) SY